Senseless TR

SenseLess: A Database Driven White Spaces Network
Rohan Murty , Ranveer Chandra, Thomas Moscibroda, Paramvir Bahl Microsoft Research, Harvard University
Microsoft Research Technical Report MSR-TR-2010-127 September 20th, 2010
Abstract
The FCC ruling on Nov 4th, 2008 on white spaces has opened up the possibility of wireless network deployments over white spaces, i.e., vacant UHF TV channels. A key requirement for any white space device (WSD) is that it must ensure that none of of its transmissions interfere with incumbents, namely TV transmitters and wireless microphones. The FCC ruling proposes two techniques for WSDs to meet these requirements: spectrum sensing and the use of a geo-location database. A host of prior work has focused on building better spectrum sensing techniques for WSDs to determine those parts of the spectrum that are currently occupied by primaries. While potentially feasible, this approach is technically challenging. Hence, in this paper we propose Sense- Less, an alternate design and approach towards building a white spaces network. As suggested by the very name, in SenseLess, WSDs rely less on spectrum sensing to determine white spaces availability. Instead, they primarily rely on a combination of an up-to-date database of incumbents, sophisticated signal propagation modeling, and an efficient content dissemination mechanism to ensure efficient, scalable, and safe white space network operation. We build, deploy, and evaluate this infrastructure service and compare our results to ground truth spectrum measurements. We present the unique system design considerations in the design and implementation of SenseLess that arise due to operating over the white spaces. We also evaluate the efficiency and scalability of SenseLess.
1 Int roduction
The recent transition from analog to digital television has opened up significant amounts of RF spectrum. The FCC has seized this opportunity, and in a recent ruling [2] has allowed unlicensed devices to opportunistically operate in unoccupied portions of this spectrum, also called white spaces. As the DTV transition begins to roll out worldwide, regulatory agencies of several countries are explor- ing similar possibilities of allowing unlicensed devices to communicate over white spaces.
The FCC ruling permits unlicensed devices to transmit in white spaces as long as they do not interfere with the primary licensed users of this spectrum, i.e. TVs and wireless microphones (mics) [2]. To prevent this interference, the ruling proposes white space devices (WSDs) must rely on a combination of local spectrum sensing and a database of incumbents to predict white space availability. In spectrum sensing, each WSD directly samples the spectrum directly to determine white space availability. The database, on the other hand, is much less fleshed it. Abstractly, it is a repository of a primaries (or incumbents) with some relevant information. While the database approach has been far less trodden, there has been plenty of prior work on spectrum sensing techniques. Such work has proposed and investigated the use of spectrum sensing to determine the presence of primary users, e.g. [34]. Since TV and MIC receivers are passive (not transmitting RF signals), white space devices need to sense the presence of a primary user at very low thresholds in order to ensure sufficient physical distance to the primary receiver and avoid causing interference. The FCC specifies this threshold to be -114 dBm.1 For a variety of reasons, sensing is not only very diffi- cult and expensive from a technical perspective, but it is also prone to false alarms, i.e., it tends to be overly conservative thereby preventing unoccupied channels from being used and causing white spaces to be wasted [37]. Incidentally, in their testing of submitted prototype devices, the FCC has reported similar findings [17]. This suggests that requiring and relying solely on sensing could become a major impediment for the proliferation of white space networks, and more generally, cognitive wireless networks that optimize spectrum efficiency using opportunistic spectrum access. Motivated by this observation, we explore the feasibility of an alternate design for white space networking in which devices do not primarily rely on spectrum sensing while still ensuring the requirement of non-interference with primaries. Our design is a significant departure from conventional white space networking architectures that are rooted in clients solely relying on spectrum sensIt is reported OFCOM in the UK is even considering a threshold below -120 dBm.
1
ing. [18, 39, 23]. We present SenseLess, an infrastructure-based white space network system in which, as suggested by its very name, WSDs do not primarily rely on sensing to determine white space availability. Instead, they primarily rely on a database (also known as a geo-location service) which uses sophisticated propagation modeling (including high-resolution terrain-data as well as TVtower- specific parameters, such as antenna-height, etc.), and to compute the white spaces at any given location. The SenseLess system then disseminates this spectrum availability information to each device in the system, thus allowing these devices to communicate without the need for sensing the spectrum. Although the basic approach appears to be simple, de- veloping a comprehensive solution in which devices rely on the geo-location service to determine white spaces availability is non-trivial. This is because our design in- troduces some unique challenges which must be over- come. TV detection: Relying solely on a geo-location service to detect TVs may result in false positives causing a loss in white spaces, and false negatives resulting in interference with TVs. The extent of these errors depends on the accuracy of RF propagation modeling, which, to the best of our knowledge, has not been quantified in the context of white space availability. Client Bootstrapping: This is a chicken-and-egg problem. Due to spatial variation [18], when a client shows up, the BS needs to know the clients location to infer the available channels at the client. However, to avoid interference with primaries, the client cannot transmit its location to the BS unless it knows the available white spaces at its own location. Location Resolution: Although the FCC requires devices to know their location before accessing the geolocation service, constantly determining location is a power intensive error-prone operation [25]. So the question is, is knowing the location of the client crucial for correct operation? And if it is, at what resolution is this information needed? We address the above challenges as follows. (1) To detect TV stations, we use a geo-location service that performs propagation modeling using terrain elevation data. We validate the white spaces predicted by this service with spectrum measurements collected by a 1900 mile drive in two states. (2) To solve the bootstrapping problem, we intelligently code the white space availability in a BSs beacon packets, and finally (3) through detailed measurements, we evaluate the impact of client location information on our network. From a systems perspective, the design of the SenseLess system is to a large extent governed by constraints that are unique to white spaces. We show that any white
space system in which, instead of sensing, devices rely on a third-party entity to learn about white space availability must follow certain design principles. Specifically, we show that in order to achieve sufficiently accurate (i.e., not overly conservative) prediction of white space availability, sophisticated and computationallyintensive models must be used. This result is derived by comparing our model results with actual measurements of UHF spectrum across Washington state. Also, to avoid wasting significant amounts of available white spaces, accurate client locations must be taken into account when determining the free channels at a base station. Finally, mobility can only be handled efficiently if the service disseminates spectrum availability information to each client at very low-latency. We design and implement SenseLess following these design principles. At the time of writing this paper, the FCC is still open to reconsidering portions of its ruling. We discuss what potential implications our results may have on these rulings. It is interesting, for example, that the current FCC rulings require clients (except disconnected clients) to occasionally access a TV band database and ask for the permissible set of operating channels. This is in addition to the requirement that every device must sense the spectrum. In comparison, we advocate utilizing the concept of a geo-location service in a much broader and more fundamental way. Instead of merely taking a hint from the TV-band database as to which channels is free, devices in a SenseLess network, WSDs must rely solely on this service to determine white space availability. Hence, we make the case that it is feasible to build a service that meets the much more stringent requirements that arise from the central role the service plays in the network architecture. Note, however, in SenseLess, as per the current FCC rulings, all WSDs are equipped with sensing capabilit ies. This feature, though not required for SenseLess, may be used to optimize the performance of the system. We make the following contributions: We demonstrate the practicality of using propagation models to accurately predict spectrum availability by comparing our model results to measured ground truth data. We derive the characteristic systems challenges that arise from removing sensing from client devices, and quantify their impact on white space utilization. We design and implement a scalable system that can efficiently enable networking over white spaces without relying on sensing. We provide extensive evaluation of our system in terms of scalability, efficiency, and safety properties.
The rest of the paper is organized as follows. We mak- ing the case for general design principles of a SenseLess-
like approach in Section 3, present the SenseLess design in Section 4, implementation details in Section 5, evaluation in Section 6, a discussion of related issues in Section 7, and finally conclude.
presence of (digital and analog) TV signals and wireless microphone signals at a very low receive power threshold of -114 dBm.3
2 Backg round & Motivation

We now provide some background on white spaces and motivate the need for SenseLess.
2.2 Sensing
The FCC proposes very low thresholds for sensing because primary users receivers, i.e. TV sets and mic receivers, are passive, and therefore cannot themselves signal the channel as busy. However, accurate sensing at these low thresholds is very challenging in practice and we aim to remove this complexity from devices primarily due to the following two reasons. First, detecting TV transmitter signals at thresholds as low as -114 dBm and across a channel bandwidth of 6 MHz is expensive. For example, most Wi-Fi cards have a noise floor above -100 dBm. This value comes mostly from the low noise amplifiers (LNAs) used in the RF cir- cuitry and from other electronic components. The noise floor of an RF receiver is given by kT + 10log(BW ) + N F [8] where BW is the sampling bandwidth and N F is the noise figure of the device. Therefore, for higher bandwidths, the noise floor is much higher. Hence, to be able to sample across a 6MHz bandwidth while detecting TV signals down to -114 dBm, we need a device with a very good noise factor and very low NF. Although one could potentially build such a device with high-end cir- cuitry, such a system would be expensive for commercial deployments. While this may change over the long term, we believe that in the near future this limitation is likely to impact the design and proliferation of white space devices. Second, there is a tradeoff between using dedicated RF frontends for sensing and the throughput achieved in such networks. There are two broad approaches for a white space device to sense a primary user: inband (sensing the primary on a channel currently being used by the white space device) and out-of-band sensing [38, 39, 29] (sensing the primary on a channel different from the one currently used by a white space device). In-band sensing does not require a dedicated RF front end but it requires periodic coordination of silence time in the network when white space nodes stop transmitting and sense for primary users [38]. Such silent periods may last between 5-72ms and need to be maintained often to ensure non-interference with primaries. Hence, in-band sensing results in lower throughput [36]. Out-of-band sensing on the other hand typically requires an RF front end dedicated towards sensing since white space nodes need to constantly scan the spectrum to detect the presence of primaries. Therefore, from a design perspective, we must either add an extra radio at the white space device that is devoted to sensing or suffer a drop in through-
2.1 White Spaces

White space is the term used by the FCC for unused TV spectrum in the VHF and UHF spectrum. There are a total of 49 TV channels in VHF (Channels 2-13; 54-216 MHz) and UHF (Channels 14-51; 470-698 MHz, except channel 37), and each channel is 6 MHz wide. Recent studies have shown that a significant fraction of these channels are actually unoccupied [30, 18, 35]. In the US, as well as in many other countries, (registered) wireless microphones are also incumbents in the UHF TV bands. Hence, as per the FCC ruling [2], unlicensed devices in white spaces can operate on a TV channel only if its transmission neither interferes with a TV transmission nor with a wireless microphone. The FCCs new regulatory rules thus open up an opportunity to develop new wireless networks to utilize this spectrum that are based on opportunistic and dynamic spectrum access. In view of the superior propagation (long range) and building penetration properties of VHF and UHF bands compared to other unlicensed spectrum bands like the 2.4 or 5 GHz bands, this has understandably created a lot of excitement in both industry [35] and academia [18, 23]. As characterized in [18], spatial and temporal variation of spectrum availability are among the fundamental differences between white spaces and other unlicensed spectrum bands such as Wi-Fi. In white spaces, the available channels not only depend on the clients location, but can also change at any instant in time (for example, if a MIC is turned on). Consequently, there needs to be a way for nodes to i) determine which parts of the spectrum are free and ii) to quickly react to changes in the availability of the currently used channel. FCC Ruling: In order to provide protection for primary users in the TV bands, the FCC ruling mandates the use of two protection mechanisms.2 First, any fixed and non-peer-to-peer portable device (so-called Mode 2 devices in FCC jargon) must access a TV band database in order to determine the available channels. In addition, however, the ruling also requires that every device operating in the TV bands must sense the spectrum for the
2
The FCC ruling in its entirety is a vast set of specific regulations.
Our description only gives a brief summary of the points most relevant to our work. For more details, we refer to [2].
3 OFCOM
is even considering a threshold of -126 dBm in the UK.
put due to silence periods. Adding an extra radio creates additional problems having to do with shielding between the radios.
3 Design Principles
A white spaces network in which WSDs rely on the rely on a database to learn of white space availability, still needs to ensure that (i) they do not interfere with primary users (safety) and (ii) it makes full use of the available white spaces (efficiency). While we describe our specific design and implementation in Sections 4 and 5, respectively, we now outline three design principles that are key to building such a system. Our principles are guided by answers to the following questions: How does the database determine white spaces availability, and how closely do the predictions match up with the ground truth? Should white space devices know their respective locations, and if so, to what resolution? And, what is the impact of mobility? The following three subsections give answers to these questions.
3.1 Determining White Spaces Availability

To learn the white space availability at a given location, the database could use one of two schemes: use spe ctrum measurements for that location, or compute spe ctrum availability using RF propagation models. The former, a data-driven approach, requires extensive wardriving measurements at low sensitivity thresholds and m ay take a long time to be complete. Furthermore, the me asurements will have to be repeated whenever the pr imary users transmission characteristics, such as tran smit power, antenna height, license terms, etc., chang e. In our experience, these changes are not uncommo n. The latter, a model-driven approach does not suffer fro m these drawbacks, and our SenseLess system takes th is approach. However, the key question of any such mode ldriven design is the validity of the RF propagation models. There exist a plethora of propagation models for UHF signals, each of varying complexity and accuracy. We use real-world measurements to compare the accuracy of different models for computing the available white spaces. Specifically, we compare the following four well-known propagation models: Free Space, Egli, and Longley-Rice (L-R) with terrain, and L-R (without terrain). We restrict ourselves to these models because as per the FCCs prior rulings [16], L-R is the default model to be used for predicting signal attenuation. We exclude detailed descriptions of these models due to lack of space. Briefly, Free Space is the simplest model and
has been used to study white space availability in a number of research papers [26, 20]. Egli is often used to compute RF propagation in frequencies between 50 and 1900 MHz. L-R is a much more complex propagation model, that takes into account climactic effects, soil conductivity, permittivity, Earths curvature, and surface refractiv- ity, and has also been used by the FCC to compute TV contours [16]. L-R with terrain also takes elevation data as input and is very computationally intensive. The question we seek to answer is, how closely does the white spaces availability predicted by these models resemble the ground truth? And, is relying solely on any of these models really a viable alternative to relying entirely on spectrum sensing? We are not aware of prior work that answers all these questions. 3.1.1 Measu rement Methodology We measured the UHF spectrum across the state of Washington. As shown in 3.1.1, spectrum measurements were taken during the months of July-August 2009, across a driving path of 1500 miles, at a set of 57 diverse locations including large cities, downtowns, suburbs, between large buildings, mountain ranges, forests, valleys, at the edge of water bodies, and also across areas of different population densities.
MHz. We then measured the channel power of the spectrum across this 6 MHz band. We mark a channel as being occupied if the corresponding chan-
Figure 1: A 1500 mile path over which we measu red UHF

spectrum at 57 different locations. The push pins indicate points of measu rement.
Using a spectrum analyzer fitted with a UHF antenna, we measured the signal strength for all 30 UHF channels (21 through 51 except channel 37, which is reserved for astronomy) by restricting the channel on the spectrum analyzer at the center frequency of every TV channel and a bandwidth of 6
LR (Terrain) LR (w/o Terrain) 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Egli Free Space
are able to achieve a satisfying accuracy; using simplistic propagation models results in a waste of white spaces. 3.1.3 More White Spaces using Variable Thresholds We note that using a geo-location service, the BS can do better. A geo-location service allows the BS to determine the source of TV transmissions as well. Since the FCC has put forth different protection contours for different
CDF
20
40
60
80
100
when measuring channel power values.
Fraction of white spaces lost (%)
Figure 2: Comparison with ground truth for diffe rent

propagation models. We lose fewer white spaces by using terrain based Longley Rice propagation.
nel power is less than or equal to -81 dBm4 , and available otherwise. Using the same threshold, we compare these findings with those predicted by different models and high resolution terrain data (measured at 100 m intervals across the planets surface). If a channel is occupied in the ground truth data but available in the models, we flag it as a false positive; if it is available in the ground truth data but not by the model, it is a false negative. 3.1.2 Comparison of Different Models In our entire data set of 57 locations and 30 channels each, none of the models gave any false positives.5 So, they all met the safety requirement. To quantify efficiency for each location, we express the number of false negative channels as a fraction of the total white spaces available at that particular location. The results are shown in Figure 3.1.2 which is a CDF of the fraction of channels lost because of false negatives. It can be seen that the median loss rate for L-R with terrain is only around 8% of the available channels. To us this was a surprising result since it suggests that with careful modeling it is possible to 1) not lose too many white spaces and at the same time 2) have very few or no false positives. In contrast, all other models result in many white spaces being wasted. Hence, we can conclude that that a purely modelbased approach based on the L-R model with terrain data can predict white space availability at a low false negative rate. In other words, the loss of white spaces by not sensing is low. However, one observation with important systems implications is that only complex and computationally-intensive models based on terrain data
4
This is a limitation imposed by our spectrum analyzer hardware
classes of digital or analog TV transmitters [2], the geolocation service can appropriately change the threshold for computing white spaces. For example, high power UHF DTV stations need a minimum signal strength of 51.0 dBu at a receiver whereas for low power UHF DTV stations it is expected to be 41 dBu [2]. There are similar such differences for analog TV transmitters as well.6 The authors in [19] recognize the need for differentiated thresholds depending on the source of TV transmissions. We build on this insight and determine the precise thresholds to be used for each class of TV transmitter. In Section 9.1.4, we evaluate this approach via ground truth measurements and our geo-location service.
3.2 Need for Location

At first glance, it is not obvious why clients in white space networks that do not primarily rely on sensing need to be equipped with location information. As suggested by prior proposals [1], and the FCC ruling [2], it should be possible for Base Stations (BSs) to
5 In the free-space model, we used a very conservative path-loss exponent of 2. Using an exponent of 3, we measured several falsepositives.
either use their own locations as a substitute for locations of clients associated with them, or to conservatively operate only on channels that are free in the entire coverage range of the BS. The intuition is as follows. Each BS can merge the white spaces availability at every single location in its coverage area to derive a comprehensive map of white spaces availability. A channel that is free everywhere in its coverage area can be used to communicate with a client that is located anywhere in its coverage area. The problem is that this method can result in a loss of available white spaces as the comprehensive map is a result of a union of all occupied channels within the coverage area of BS. In this section, we quantify this loss (and thus also the need for client-location). Methodology: We use L-R with terrain data (see Section 4) to compute RF propagation from TV towers. We emulated the deployment of a BS in 90 points 30 each of city downtowns, towns and suburbs in the US. We fixed the transmit power to 4W EIRP, as specified in the FCC ruling and computed the coverage area of the
The FCC has granted a longer transition period to certain classes of analog TVs to remain on the air.
6
Cities Suburbs Towns 0 20 40 60 80 100
0 20 40 60 80 100 % UHF Channels Lost
% UHF Channels Lost
White Space Loss(%)
1 Cities 0.75 Suburbs 0.5 Towns 0.25 0
1 0.75 0.5 0.25 0
100 80 60 40 20 0 0 0.5 1 1.5 2 2.5 3 3.5 4 Location error (miles)
CDF
(a) 4W BS
CDF
(b) 100mW BS
% UHF Channels Lost
60 50 40 30 20 10 0 Cities Suburbs Towns 0 0.5 1 1.5 2 2.5 3 3.5 4 Transmit Power (W)
(c) Varying Transmit Power
BS that receive better signal reception from TV transmitters than other points. Hence, the white spaces availability in such
Figure 3: The impact of using BS location as a proxy for

client location on the white spaces lost, as a function of the BS transmit power.
BS (again using L-R with terrain data model, see Section 4). We divided the coverage area in 100m100m miles grid, and for every grid point we computed the available white spaces. We declared a TV channel to be available at a grid point, if the receive signal strength from any TV tower is not greater than -114 dBm, which is the sensing threshold specified in the FCC ruling. For any BS location, the number of white space channels lost is computed as (C Loc CnoLoc )/C Loc , where CLoc and CnoLoc represent the number of available channels with and without location, respectively. CnoLoc is the number of channels that are available everywhere in the BSs coverage area: these are the channels a BS can select without having client location information (because the BS must assume that a client could be anywhere within its coverage range). In contrast, CLoc is the number of common channels available between the BS and a client that is located randomly at one of the grid points. These are the number of channels the BS can select if it knows its clients location. White Spaces Lost Without Location: In Figure 3 we plot the CDF for the number of channels lost across the 90 locations. The median loss of white spaces by using the BSs location as a proxy for client locations results is more than 80%. There are two reasons for this high loss. First, without knowing client location, the BS is forced to be very conservative and use only those channels that are available even in the most distant area of its coverage area. Second, due to changes in terrain profile, there are often points in the coverage area of the
Figure 4: The impact of location errors on the loss of white

spaces.
locations is lower than other surrounding points.7 We conclude that clients must know their locations for SenseLess to utilize the available white spaces efficiently. Location Granularity: A remaining question is to what resolution clients must know their location. GPS is accurate outdoors, but does not work in indoors. Cellu- lar technologies can yield location errors of more than a mile [32]. Location errors are problematic as they can manifest themselves in the form of both false negatives and (even worse) false positives. In order to avoid false positives, an additional safety protection range would have to be used by the system, which would result in an additional loss of white spaces. To quantify the impact of location errors, we measure this loss in available white spaces. We show results for the worst of the 90 locations. As Figure 4 demonstrates, higher location errors causes significant reduction in the overall white spaces. For example, an error of 2.5 miles can result in 80% loss of white spaces. We conclude that for a system without sensing at white space devices to function effectively, clients must have location information within 0.5 miles.
3.3 Handling Mobility

In SenseLess every WSD primarily relies on the database to determine the white space availability. Hence, this results in a delay in the WSD learning about changes in spectrum availability. Either the WSD will have to poll the database, or the database will have to push updates to the WSD. This problem becomes worse when the devices are mobile. If mobile, the WSD could have travelled some distance between the time it receives two subsequent spectrum updates (Figure 3.3). To address this problem, mobile devices will have to add a protection range of d to determine the white spaces that it can use. That is, even though a channel might be available at a devices location, it must not use it, if this channel is blocked off at any location within distance d of
7 For this reason, a similar experiment using inaccurate propagation model without terrain data would result in a much smaller loss of whitespaces due to not knowing location.
SenseLess Service
Back-end Store
Mobile Node
Primary
Engine
BS
BS
Figure 5: Impact of mobility on the protection region

around an incumbent. The extra d is requi red to account for delay in informing the BS and client about the presence of a new incumbent.
(a) SenseLess architecture overview

Back-end Store
Terrain Servers
the current location. For example, for a polling interval of 1 minute, a mobile client that accurately knows its location and can travel at 60 mph will have to add a protection range of 1 mile. As we see in Figure 4, this would result in 20% loss of available white space spectrum. dle mobility, a white space network that relies on the database to determine white space availability must disseminate spectrum changes to the clients at very low latencies.
Primaries db
SenseLess Engine gation Models ... Longley Rice
Mic Registrar
In this section, we have derived characteristic system design constraints and challenges for a system that does not perform sensing at the white space devices: The system must be driven by highly-sophisticated signal propagation models that are based on terrain data. Such models are computationally expensive. Using conservative estimates for clients with un- known location results in a significant waste of white spaces, i.e., client location must be incorporated into the design. In order to handle mobility, information dissemination latency must be very low because otherwise, a significant amount of white space is wasted. Besides these design constraints, there is one more challenge that needs to be solved. Since clients lack sensing capabilities, there must be a mechanism for the system to learn about the existence of wireless mics. In the next section, we show how our system, SenseLess, addresses all the above constraints and challenges.
4 SenseLess: System Design

In this section we describe the design of SenseLess, details of its internal components, and how SenseLess adSenseLess Service
API
(b) Internal components
Figure 6: The SenseLess System Architectu re dresses the challenges derived in the previous section.
4.1 System Overview

Figure 6 illustrates the overall SenseLess architecture and its internal components. The aim of the system is to enable infrastructure-based wireless networks operating in white spaces that primarily rely on the database as a means to determine white space availability. At the heart of the system lies the SenseLess Service,
a logically centralized entity to which client devices and base stations (BS) are connected to. The SenseLess service determines, for any given location, which parts of the spectrum are white spaces. Any change in spectrum occupancy (e.g. due to say a TV transmitter being switched off at night or a wireless microphone being switched on) is reflected in the SenseLess service. All WSDs, be they clients or BSs, rely solely on the
API G hiteSpaces( etW L) G rim etP aries( L) U pdatePrim ary( p) R g r( e iste n, L) CoverageArea( b) SetC over( b)
Description Returns white spaces at location L Returns list of primaries and RSSI at L Update information on a primary p Register BS/client n with grid location L Computes coverage area for a BS b Computes set cover channels for coverage area of BS b
We intend to open up the service for public use by NSDI 2010.
Table 1: API exposed by the SenseLess service.
SenseLess service to learn about the availability of white spaces at their respective locations.
4.2 SenseLess Service

The primary task of the SenseLess service8 is to respond with an accurate white spaces availability for a given location. The white spaces availability is a bitmap of channel availability at a given location, where location is rep- resented by L = L atitu e, d Longitude . The APIs provided by the service is shown in Table 1. The SenseLess service can operate in two modes. First, it can be queried giving a specific location as input. Alteratively, using a publish-subscribe model, the SenseLess service can track changes in the white spaces availability of a BS or its associated clients. On detecting changes in the white spaces availability at any of these locations, the service fires an event that informs the BS of the changes in white spaces availability either at the BS itself or at one its clients. Hence, to support both modes, an active connection is maintained between the service and each BS. The BS then disseminates this spectrum availability information to its clients. As demonstrated in the previous section, the accuracy of the spectrum map depends on the quality of the propagation model and fine granularity of the terrain data. Therefore, the SenseLess service is comprised of two components: the back-end store that maintains the terrain and primary user data (such as TVs and wireless microphones) and the SenseLess engine that accurately computes the white spaces availability at a location. These components are illustrated in Figure 6(b).
4.3 Back-end Store

The back-end store consists of the database and the terrain server. Database: The DB stores the following information: Information on all television transmitters including the TV tower location, channel, height, transmit power, antenna directionality, mechanical beam tilt, etc., all of which are updated on a daily basis. This
is based on publicly available data from the FCCs Consolidated Database System (CDBS) [24] Wireless microphone registration data including location of wireless microphone, channel, and time of last report. We discuss how this data is obtained in Section 4.5. The database also serves as a cache for the white spaces availability computed previously for various locations. We explain this further in Section 4.4.1. Terrain Servers: The primary responsibility of the terrain server is to provide terrain elevation data at any arbitrary point on the planets surface. It stores high resolution worldwide terrain elevation data. We obtain this data from various publicly available government efforts for mapping the planets terrain. We use various elevation data sets, such as GLOBE [14], measured at 1 kilometer intervals across the planets surface, or the high resolution SRTM data [11] from NASA measured at 30 meter intervals. The terrain server provides terrain elevation data between the two points at a specified interval. To this effect, the terrain server receives requests specify- ing (LT X , LRX , Interval) where LT X and LRX are the coordinates for the transmitter and receiver, respectively, and interval is the resolution at which terrain data must be sampled between these two points. Given this input, the server computes a direct path between the LT X and RT X along a great circle, and returns elevation samples between the two points at the specified interval.
200ft, and at a frequency of 515Mhz (lowest channel), that range can be computed to be roughly 300 miles.
4.4 SenseLess Engine

To accurately determine the white spaces availability at a location L, the SenseLess engine computes the attenuation of UHF signals via propagation modeling by performing the following steps. 1. List all primary users transmitters within a large search radius from L. Conservatively, these are all the primary users whose signals could potentially be heard at L at an RSSI greater than the specified threshold.9 2. If the primary is a TV transmitter, the engine fetches the elevation data between the TV tower and L from the terrain server. If the primary is a mic, we assume a conservative fixed protection radius around it ([2]). 3. The elevation data in conjunction with the transmitters parameters (such as height, power, antenna directionality, etc.) are used to determine the signal attenuation using the Longley-Rice (L-R) propagation model. The attenuation in turn is used to compute the RSSI of the transmitter at L.
9
For TV towers with 1MW Tx power, an antenna height of
1 0.75 0.5 0.25 0
30
60
90
120 150
the channel in the white spaces availability of the affected grid points and recompute only the affected points. In re- sponse to queries for a given location, the SenseLess En- gine returns the precomputed spectrum map of the near- est grid point. We evaluate the efficacy of this approach in Section 6. Per-Region Terrain Server: The cost of fetching terrain data from the back-end terrain servers is high since such requests go over the network and read from disk. Consequently, we maintain a per-region terrain servers that store an in-memory copy of the regions terrain. As we show in Section 6, this optimization significantly reduces computation latency. We also maintain a terrain cache on the SenseLess Engine that corresponds to the latest tile fetched from the terrain server.
CDF
Interval Between Changes (days)
Figure 7: Changes to entries in the FCC TV transmitter

database over a period of three months (monitored on a day-to-day basis).
4. The white spaces availability at location L is created. As part of this, a channel C is deemed to be blocked if there is at least one primary who is active, i.e., an incumbent whose RSSI at L is greater than -114dBm. 1 if an incumbent is active on C at L M ap L (C ) = 0 otherwise
4.5 Detecting Microphones

In SenseLess, we support wireless microphones using incumbents whose signals are received at -114 dbm at the center of the grid. Note, events such as changes in a TV transmitters configuration can impact some grid points. To efficiently handle such events, we invalidate
The critical step in this process is determining the attenuation of a UHF signal over a certain distance and how this affects the RSSI. This is achieved via propagation modeling of UHF signals using the Longley-Rice propagation model, as discussed in Section 3. We use the same parameters for L-R propagation as those specified by the FCC [16]. 4.4.1 Improving Performance The SenseLess engine plays a crucial role since it computes L-R propagation using terrain elevation data. This particular task is computationally intensive and can impact the latency of the results returned. To improve this latency, we make two key optimizations to the engine. Precomputations: We monitored the FCCs TV tower database (updated daily) over a period of three months and found additions/deletions/changes to licenses were made on average, once in 2 days (Figure 7), which is relatively infrequent when compared to wireless microphones being switched on/off several time a each. Also, as seen in Figure 7, of the 10,000+ entries in the FCC database on average, fewer than 1% of the entries are changed, and we can quickly recompute those grids only these few entries cause. Hence, we note that once we compute the white spaces availability for a location, in most cases, we can cache and reuse this information for future queries. We enable this optimization and cache the white spaces availability for each grid point in our back- end database. We divide the planets surface into grids of size 100m100m. For each grid we store the list of
one of the following techniques. Manual updates: We provide an API in SenseLess to add an entry for a wireless microphone as an incumbent user. Authorized users can add an entry for the microphones frequency channel, transmit power and the location and duration of the event. SenseLess then treats the wireless microphone transmitter as a (very) low power TV tower for the specified duration. These entries can be expired after a time out period. MicSentry: We provide an alternative to manually updating the database. Our technique consists of a small device, which we call the MicSentry, that is plugged in close to the wireless microphone receiver. The Mic- Sentry detects the presence of the wireless microphone and updates the SenseLess back-end database using the UpdatePrim ary API described in Table 1. It uses an alternative technology, such as 3G, Wi-Fi, Ethernet, etc, to connect to the SenseLess service. When the MicSen- try does not detect the wireless microphone for a prede- fined amount of time (in our system, 5 minutes), the entry is deleted from the database. We also propose and implement an alternate version of the MicSentry which does not require a direct connection to the database. Instead, it acts as an amplifier for the wireless microphone and signals the mics presence via beacon packets. This is described in further detail in Section 9.
4.6 System in Operation

During the operation of the system, each BS must learn about which channels are available at itself and its clients in order to select an feasible white space channel. We now show how this information is conveyed to a BS.
Every client periodically sends its location and the technology used to determine the location to its associated Base Station. The BS uses this information to determine the commonly available white spaces at the client and itself. From among all these usable white spaces, the BS selects an appropriate chunk, and clients associate to the BS on this channel.10 The BS also subscibes to push updates from the SenseLess Service at the clients location as well as grid points around it depending on the location error of the technology used to determine the location. We use conservative values for the location error, e.g. 50 meters for GPS, 150 meters for Wi-Fi and 1 mile for GSM based localization. Thus, SenseLess requires all clients and BSs know their location. This design choice is motivated by our findings in Section 3. In SenseLess, location can be determined by any means, including existing technologies based, e.g. GPS, Wi-Fi [12] or GSM [32]. In the future, we plan to use TV technology for localization [9]. The above approach works once all clients are associated to the BS and as long as white space availability does not change. We will now explain how SenseLess handles bootstrapping and spectrum updates. 4.6.1 Bootstrapping Bootstrapping is a critical problem in any SenseLesslike system in which clients rely on the database to determine white space availability. When a new client arrives, it cannot transmit any packets since it does not know the white spaces availability at its location. Therefore, it is unable to even transmit its location to the BS since the corresponding packet could interfere with a primary user nearby. This bootstrapping problem would not occur if the BS always transmitted on a channel that is available at every location in its coverage area. In that case, a client could simply discover the BS and associate to it. While this would greatly simply our design, our results in Section 3.2 show that 1) a commonly available channel across the entire coverage area does not exist in most places, and 2) even in places where such a channel exists, using only this channel would result in a significant loss of white spaces. Hence, the channel selection at the BS in SenseLess takes into account actual client locations, which results in bootstrapping becoming a challenge. Channel Availability Beacons: To resolve the dead lock during the bootstrapping phase, every BS periodically (once every second) broadcasts a beacon containing the channel availability in all regions of its coverage area.
10 In this paper, we only consider the problem of determining the set of available channels. The problem of deciding which spectrum chunk among the available channels should be used is orthogonal and has
been studied in prior work [18, 23].
Specifically, a beacon contains for each grid-point in its coverage area one available channel that a client located at that grid-point could use to contact the BS.11 At a lower frequency (once every minute), the BS switches to each of these channels and listens for clients that want to join. The switching overhead can be reduced by using techniques presented in [18]. When a client joins the system, it listens for beacons from the BS and moves to the channel that is marked in the beacon as being available in the grid pertaining to the clients current location. The client then broadcasts its location on this channel, which is eventually picked up by the BS. Hence, the beacon is used as a lookup table by client when bootstrapping. Algorithm for Reducing Beacon Size: A challenge when designing the above system is to efficiently communicate the channel availability for all regions in the BSs coverage area. Transmitting one channel for every grid-point in the coverage area of a BS will result in prohibitively large beacons. For instance, at a coverage range of 10 miles, and a grid-granularity of 100m, the beacon size would be more than 100 KB, if 5 bits are used to encode a channel number. To reduce the beacon size, the BS performs the following algorithm. First, in the initialization phase, it queries the SenseLess engine to compute its potential coverage area (PCA). The PCA is defined as the set of grid points in 100m100m steps where the receive signal strength from the base station is greater than -90 dBm, i.e. the receive sensitivity at the lowest rate for many wireless cards. The PCA is computed using the L-R model. For every grid point within its PCA, the BS then retrieves the set of available channels.12 During the operation of the network, the BS uses this information to compute a minimum channel cover (MCC). An MCC for a potential coverage area A is a set of channels of minimum cardinality, such that for every grid point location within A, there is at least one available white space channel in in the channel cover. Formally, let C and LA be the set of channels and location in the potential coverage area A. A channel cover M C CA is defined as a minimum-sized set M C CA C, such that for every L LA , it holds that |C M C CA andM ap L (C ) = 0| 1. In SenseLess, we use a standard greedy set-cover approximation algorithm to compute a good approximation to the MCC. The BS now knows that at least one of the channels in MCC is available at all grid points in its coverage area.
11 Notice that when estimating the coverage area of a BS, we can be conservative and assume a large potential area. Because spectrum availability computation at the BS takes into account, this does not have negative consequences on the amount of white spaces that can be used, it merely increases the beacon size. 12 Note that this is a one-time step and is performed only when the Base Station is first set up. Subsequently, the SenseLess engine only
sends updates.
The beacon contains the following information. The set of channels included in computed M C CA .
the
reason for our system to adopt a push-based model as otherwise, client-devices would have to poll the database at
A listing of one available channel from among the channels in M C CA for every grid point in its coverage area (given in row major form). Each channel is encoded using dlog(|M C CA |)e bits. As we show in Table 4 of our evaluation in Section 6, the size of M C CA is at most 4 in the US, and hence, 2 bits per grid-location suffices. Finally, a BS uses RLE compression to compress M C CA information about adjacent grids where the channel availability is similar. This significantly reduces the size of the beacon. We evaluate this approach in Section 6.3. 4.6.2 Spectrum Updates To handle updates in spectrum availability, such as the appearance of a mic, a BS subscribes for push updates from the SenseLess engine for all grid points in its coverage areas. Therefore, when a mic is switched on and the SenseLess engine receives notifiction of this event, it first quickly determines those grids impacted by the mic. We show in Section 6, the time taken to do this is very low (less than 500 ms when supporting up to 1000 mics being concurrently switched on). Based on this information, the SenseLess engine determines if a mic is switched on in the coverage area of a BS. If that is the case, the BS is notified of those grids that are impacted by this change in the white spaces availability. This is done automatically as part of computing the PCA, as described above. When either a new primary user appears, or the parameters of an existing primary user change, the SenseLess engine computes the changes to the affected grid points around the primary user and sends the updates to all the BSs that have subscribed to changes for these grid points. If needed, the BS recomputes the spectrum it is operating over and/or the contents of the beacon. We note that the above steps affect the latency in propagating a change in the available white spaces. We evaluate the latency in Section 6. We note that when a primary user is already active, our push-based system architecture is very nimble in handling client mobility. Since the BS has an up-to-date view of spectrum availability and the client locations, it can precompute and adapt in ad- vance to the spectrum that is available at all its clients, thereby avoiding client disconnections as well as interference with the primary users. As we have shown in Section 3, keeping query latencies low is a critical design constraint in white space networks when supporting mobility. This is the prime
very high frequencies.13
5 SenseLess: Implementation
We have implemented a prototype of the SenseLess that includes the service, engine, and the back-end store. The entire system is implemented using C# and consists of approximately 13000 lines of code. Back-end Store: We use SQL Server to implement the database in the back-end. This database stores primary user data as well as precomputed white spaces availability for various locations (Section 4.4.1). We implement per-region terrain servers as multithreaded processes that store a large in memory cache of the terrain data. We run this process on server class machines with at least 8GB of RAM. Hence, we limit the in- memory cache for a terrain server to 4GB. We have also implemented a terrain cleaner process that periodically removes untouched or old elevation tiles from memory. SenseLess engine: The SenseLess engine is implemented as a multi-threaded service capable of accept- ing multiple requests for either white space availability or updates to the back-end databases. As part of the engine, we have also implemented a library of various propagation models, including the point-to-point variant of L-R [15] and a broadcast version that does not factor terrain data L-R (no terrain). We have also implemented other popular and simpler UHF propagation models such as Okumura-Hata, Egli, and Free Space. All propagation models are implemented as dynamic-link libraries that can be loaded/unloaded easily at run time. BSs access the SenseLess service remotely via the SenseLess service. We divide the entire US into grids of area 100 sq. m since this is close to the smallest possible resolution supported by our elevation data. Our investigations reveal white spaces availability due to TV transmitters tend remain more or less uniform within the grid. The local terrain cache in the SenseLess engine is a 10 MB cache that stores the recently fetched terrain data from the terrain server. SenseLess service: BSs and clients to interface with the SenseLess via the SenseLess service. We have two implementations of the service: (i) as a web service and (ii) as a multi-threaded socket based server. While we use the latter for all our experiments, we consume the web service to build a front-end web based GUI for Sense- Less, shown in Figure 8. Terrain Data: We obtain the terrain elevation data from publicly available high resolution SRTM4 and GLOBE data sources, which is maintained by NASA.
13 In fact, sensing-based white space devices have the same drawback; they will have to sense (=pull) at very short intervals in order to sustain client mobility. Thus, the FCC ruling mandates client devices to sense every 30s.
Additional channels gained 0 1 2
Number of Locations 2 10 3
Table 2: Ground truth measu rements of additional white

spaces gained by variable thresholds across 15 location in Massachusetts.
15 12 9 6 3 0
Count
Cities Suburbs Towns
Figure 8: Screenshot of a web based front-end GUI for

SenseLess showing the incumbents as push pins on a map, a table of the list of nearby primaries and their respective RSSI, and white spaces availability at the specified location.
Number of additional channels gained
Figure 9: Predicted gain in white Spaces from variable

thresholds.
The SRTM4 dataset contains elevation data measured from satellites at approximately 100 meter intervals on the planets surface while the GLOBE data is measured in 1 km intervals. Adaptive Elevation Selection: Due to errors in measurement and conversion, it is common to find holes in the elevation data sources. These are more prevalent in higher resolution data such as SRTM than in GLOBE, which is a low resolution data source. Hence, governmental agencies responsible for the data take additional steps to fill in these holes via a variety of techniques such as bilinear interpolation or gradation smoothing. However, in spite of such efforts, anomalies in the terrain data continue to exist, which in turn, impacts the results of the terrain based propagation models. For example, our L-R implementation returns an error code signalling errors in which parameters are out of range. We have found, these are typically caused by errors in the elevation data. To counter such anomalies, we implement an adaptive elevation switching strategy whereby for a signal propagation computation between two points, we first try the higher resolution elevation data. If the propagation model returns an error stating some parameters are out of bounds, we then examine the path profile between the two points to determine if there are holes in the path. If that is the case, we then switch to the lower resolution GLOBE data and try the same again. If that fails as well, we finally switch to the L-R (no terrain), an L-R variant that computes signal propagation without taking elevation data into account (as opposed to point-to-point which factors in elevation data between two points). As we will demonstrate in Section 6.2, this strategy significantly improves the accuracy of the results by reducing false positives and false negatives.
6 Evaluation
We evaluate three aspects of SenseLess: (i) effectiveness in handling holes in the terrain elevation data, (ii) efficacy of our beacon reduction technique, and (iii) system scalability. In our experiments the SenseLess Engine and backend databases ran on a single server class 64-bit quadcore machine with 8GB of RAM. We dedicated a separate machine with 4GB of RAM as a terrain server. To emulate BSs and clients, we used a rack of 4 Blade work- stations, each equipped with 4GB of RAM. All nodes were connected to each other using a 100 Mbps subnet. All measurements and evaluations of SenseLess are performed across 33 cities, 30 suburbs, and 27 towns. We picked these locations based on populations distribu- tions as advertised in various mainstream media such as CNN, Economist, Yahoo Travel, and BusinessWeek as being the most liveable places in the US. For each one of these locations, using SenseLess we deployed a BS for our experiments. Hence, in essence, we had 90 BSs. We set the default EIRP of a BS to 4W, the limit allowed by the FCC ruling.
6.1 Variable Thresholds

As described earlier in Section 3.1.3, a priori knowledge of a TV transmitter type can be used to set different detection thresholds and potentially open up more white spaces. We evaluate the benefits of this approach via (i) outdoor measurements of TV signal reception at 15 different locations in Massachusetts and (ii) predicting TV reception at the 105 locations using the geo-location service. To the best of our knowledge, this is the first such
Elevation Source Adaptive SRTM only GLOBE only
# False Positives 0 21 25
Category Cities Suburbs Towns
Channels Needed 4 4 3
Table 3: Impact of various elevation sources on the false

positives in SenseLess when compared to the ground truth measu rements.
Table 4: The minimum number of channels needed to

cover the entire coverage area for a white spaces base station deployed in 90 different cities in the US.
empirical and modeling driven study to show the benefits of variable thresholds. Using the TV contour and co-channel interference cri- teria provided in the FCC ruling [2], we calculated the thresholds for the different classes of TV transmitters. Using a hardware platform similar to the one described in [31], we measured TV reception on UHF channels (21-51) across a 400+ mile driving path in the state of Massachusetts14 . For each channel, we measured TV reception down to a threshold of -114 dBm, the minimum specified sensing threshold for TVs as proposed by the FCC [2]. For each location L, we use the geo-location service to determine the type of the TV transmitter t we expect to detect on a particular channel. From this we determine T hresh t , the appropriate threshold a TV station of type t. If the geo-location service determines two (or more) TV transmitters to be active on the same channel at L, we L pick the lower threshold. If RSSI t , the measured RSSI at L from a TV transmitter t, meets the condition:
t 114dBm RSSIL T hresh t , we classify t to be
1 0.8 0.6 0.4 0.2 0
CDF
Cities Suburbs Towns 0 20 40 60 80 100 120 140 160 Beacon Size (Bytes)
Figure 10: Size of compressed beacons as computed across

various locations in the US
found all these approaches yield a similar number of false negatives. However, as seen in Table 3, only the adaptive elevation technique results in zero false positives. Hence, it is important to adaptively switch between elevation sources when modelling.
6.3 Reducing Beacon Sizes
As discussed in Section 4.6.1, BSs need to beacon specnumber of false positives and false negatives with measured ground truth data (Section 3.1) for two cases: (i) when using L-R with only a single elevation source, and (ii) using L-R with our adaptive elevation technique. We
14
inactive at L. If a channel has no active TV transmitters left, we declare such a channel as being a white space. The results from our empirical study using this approach shown in in Table 2 demonstrate that on average we gained at least 1 channel across most locations, which is on average 50% more white spaces at these locations. We conducted a similar study across the 105 locations using only our geo-location service. The results presented in Figure 9 show that by using variable thresholds, we gain white spaces across all locales. Hence, there are practical realizable benefits for variable thresholds.
We elide a map of the driving path due to a lack of space
6.2 Countering Holes in Elevation Data

As outlined in Section 5, holes in terrain elevation data can cause false positives, i.e. claim the channel to be free when there is a primary user. This is in violation to the safety requirement of white space systems. To show the effectiveness of our approach (Section 5) in filling the holes and resolving this problem, we compared the
trum availability in their coverage area. We first measured the value of M C CA , i.e. minimum number of channels, such that at least one of them is available at any point in the BSs coverage area. As seen in Table 4, the worst case M C CA is 4. Therefore, 2 bits are sufficient to encode the channel availability at a grid point in the coverage area of a BS. We then measured the efficacy of our beacon compression approach over uncompressed beacons. Figure 10 shows our technique results in a median beacon size of 40 bytes for towns, 70 bytes in suburbs and 85 bytes in cities. We also measured the reduction in beacon size due to our compression algorithm, and present the results in Figure 11. As seen in the figure, our compression algorithm reduces the beacon size in more than 50% of cities and suburbs by more than 62%. We note that we do not see as much benefit in towns since most towns we measured at have irregular terrain and hence, BSs in these areas have a smaller coverage area. Hence, as seen in Figure 10, the beacon size in towns is smaller as well.
6.4 System Scalability

We now evaluate two optimizations that improve the scalability of our system: (i) Using the terrain servers
Time to fetch white spaces availability (s)
1 0.8 0.6 0.4 0.2 0
CDF
Cities Suburbs Towns
20 16 12 8 4 0 0 100 200 300 400 500 600 700 Number of concurrent requests
25 30 35 40 45 50 55 60 65 Reduction in Beacon Size (%)
Figure 11: CDF of reduction in beacon sizes due to beacon

compression at different locations. Beacon compression can reduce beacon sizes by as much as approximately 62% in the median case.
Figure 14: Time taken to fetch white spaces availability

across a varying number of concur rent requests, with precomputing enabled (25th & 75th percentile error bars).
Time to compute updates (s) 3 2 1 0 1 10 100 1000 10000
100 Time to Compute White Spaces (s) 80 60 40 20 0
No Terrain Server 1 Terrain Server 2 Terrain Servers
Number of concurrently active mics

0 10 20 30 40 50 60 70 Number of Concurrent Requests
Figure 15: Update latency in SenseLess: The time taken

to compute changes caused by wireless mics turning on as a function of the number of concur rently active mics. The x-axis is in log scale.
Figure 12: Impact of using Terrain servers on the time

taken to compute white spaces availability across the state of WA, as the number of concur rent requests vary (25th & 75th percentile error bars).
1 0.75 CDF 0.5 0.25 0 0 2 w/o Pre-computing with Pre-computing 4 6 8 10 12
Fetch time (s)
Figure 13: Time taken to compute white spaces availability

across 384 different locations using 32 concur rent requests, with and without precomputin g.
Precomputations: Precomputations help the SenseLess to quickly retrieve the white space availability for a given location. To quantify this benefit, we computed the time taken by SenseLess to compute the white space availability at 384 locations in WA, with and without precomputations. Terrain cache was enabled in these experiments. We sent requests for these locations over 32 parallel threads. Figure 13 shows the CDF of the time to fetch the white space availability in these locations. The than 50%. Adding per-region terrain servers further reduces the time to compute white spaces.
cache, and (ii) Pre-computing the white spaces, and then evaluate the scalability of our scheme for propagating spectrum updates (Section 4.6.2). Terrain Server Cache: To quantify the benefit of the terrain servers cache, we measured the time taken by the SenseLess Engine to compute white space availability at randomly chosen points in the state of Washington, with and without the terrain servers cache. We also var- ied the number of simultaneous requests that were sent to the SenseLess Engine, and plot the results in Figure 12. As we see in the Figure, increasing the number of simultaneous requests increases the latency in computing all the white spaces. Adding caching at a terrain server reduces this latency by more
median time to retrieve the white space availability reduces by more than an order of magnitude when using precomputed white space availability. We also evaluated this time when varying the system load, and present the results in Figure 14. We used 4 blade servers to issue the concurrent requests. As we see in the Figure, the mean time to compute white space availability for even 684 parallel requests is 10 seconds. As discussed in Section 3, for a device moving at 60 mph, this corresponds to a extra protection range of 0.1 mile, which we showed in Figure 4 to cause only few white spaces to be lost. Propagating Spectrum Updates: An important as- pect of scalability for SenseLess, is its ability to compute those grids impacted by a wireless microphone and quickly push updates about these grids to those BSs and clients that subscribe to these grids. In this section we evaluate the fast path we have implemented in the system to detect such changes to microphones. For a BS in a city, we introduced a varying number of MICs in the vicinity of the BS. We focus on the additional latency in SenseLess caused by (i) computing the specific grids
impacted by the MICs, (ii) determining the set of BSs impacted by changes to these grids, and (iii) queuing a multicast message down to the network. Note, we do not factor in the time taken for the actual multicast since a host of prior work have explored efficient ways of large scale dissemination of information, e.g. [33, 28, 22]. We repeat the experiment by increasing the number of MICs. As shown in Figure 15, the SenseLess engine can support 1000 concurrently active MICs with a median compute time less than 500 ms and 10000 MICs at a median compute time of 2.5 seconds. Note, even if we assume a worst-case-like scenario of 10 seconds to complete the multicast to inform the necessary BSs of the changes to the corresponding grids, we still only require an extra protection range of 0.1 miles with respect to a user traveling at 60 mph.
for example TV tower localization technology, e.g. ROSUM [9].
7 Discussion
We now discuss some features and potential limitations of our design, as well as future work. Impact on FCC Ruling: Based on our results, we have shown that the L-R propagation model with terrain data is reasonably accurate in modeling white space availability. Furthermore, by actively polling our geolocation service, clients can get updates in white space availability within a few seconds. Therefore, it is feasible to build white space devices by relying on the database as a primary means of detecting the white spaces available. However, for such a system to not lose out on white spaces, we do require clients to know their location within 1 mile accuracy, which supports the FCCs requirement that each client device must know (or learn) its location to operate in white spaces. Sensing: The FCCs current rulings specify all WSD devices must use the database and sensing capabilities to determine channel availability. While this paper demonstrates the operation of a white spaces network that relies solely on the database, the added sensing capabilities may be harnessed to improve the performance of the database. For example, sensing measurements from devices can be sent back to the database and cached for future lookups. While possible, this approach is technically challenging since measurements are sensitive to factors such as location and antenna. Determining Location: As described in Section 3, it is important to know the location of the devices within a few hundred meters; otherwise too much white space has to conservatively be blocked off by the service to ensure protection of primary users. Currently, we use GPS to determine the location. However, we realize that GPS is power-hungry and does not work indoors. To solve this problem, we plan to use other approaches,
Such a solution neither requires extra radios/antennas nor does it consume large amounts of energy. Disconnected Operation: Our current implementation of SenseLess works when the clients and BSs can connect to the service over the Internet. However, in some scenarios, such as in peer-to-peer ad-hoc mode, or in the worst case of Internet failure, this connectivity might not be available. We are currently working on extending SenseLess to work even when the clients cannot connect to the Internet. Our approach is to have clients carry a compressed form of the precomputed spectrum map for the region they are located. This can be downloaded when they have connectivity, say at night.
device. In this paper, we present an alternative technique, called SenseLess, that leverages propagation modeling
8 Related Work
The FCC has mandated a geo-location service and consequently several Internet websites offer the ability to view the TV coverage at a given location [3]. Spectrum Bridge recently released a web site, called Show My White Space [10], that shows the white spaces available at a given location. Although, we support a similar front end to SenseLess it differs in several key aspects. First, the focus of SenseLess extends beyond discover- ing nearby primaries. SenseLess is a complete system for enabling white spaces networking without relying on sensing. This also includes a set of design principles we espouse in our system. SenseLess also exposes a set of APIs (Table 1) that are not exposed by these services. Second, to the best of our knowledge, we are unaware if our elevation switching strategy is adopted by these other services, which we have shown as being crucial to reduce false positives and negatives. The closest work to SenseLess was proposed by Gurney et. al. [27] which presented the design of a geolocation database for white spaces networking. Their system computed the available white spaces based on the transmit power of the white space device and without using terrain information. They presented a limited evaluation in the Chicago area. In contrast, we use a more accurate propagation model that uses terrain data (Sec- tion 3). We consider various challenges in interfacing with white space client devices that arise due to location inaccuracy, bootstrapping and client mobility. We also provide a thorough evaluation of SenseLess.
9 Conclusions
White spaces offer a new frontier for wireless communication with unlicensed spectrum that has very favorable propagation characteristics. However, prior work to utilize this spectrum has relied entirely on accurate sensing of primary users at every client
at a server over the Internet to determine white space availability. As part of our system, we demonstrate the need for good propagation models, accurate knowledge of client location and quick dissemination of spectrum updates. SenseLess addresses all these challenges using a sophisticated propagation model that uses fine-grained terrain data.
http://www.fcc.gov/mb/video/tvq.html.
Refe rences
[1] Authorized Ex Parte Contact - Unlicensed Operation in the TV Broadcast Bands (ET Docket No. 04-186), FCC. [2] FCC Adopts Rule For Unlicensed Use Of Television White Spaces, http://www.fcc.gov/. [3] http://tvfool.com/. [4] Mitigating the Effects of Unlicensed Devices on Wireless Micro- phones, Ahren Hartman and Edgar Reihl SHURE Incorporated, SHURE Incorporated. [5] Pastor Joel Olsteen Voices Opposition to White Spaces Use, Radio receiver noise floor, http://www.radioelectronics.com/info/receivers/sensitivity/noise floor.php. [6] Private communication with wireless mic operators. [7] Qualnet 4.5 simulator, http://www.scalablenetworks.com. [8] Radio receiver noise floor, http://www.radioelectronics.com/info/receivers/sensitivity/noise floor.php. [9] Rosum corporation announces successful dvb-h positioning trial in collaboration with uks national grid wireless, http://www.rosum.com/pdfs/rosum nationalgridwireless dvbh technologytrial.pdf. [10] Show My White Space, http://showmywhitespace.com/. [11] Shuttle Radar Topograph Mission (SRTM), http://www2.jpl.nasa.gov/srtm/. [12] Skyhook Wireless, http://skyhookwireless.com/. [13] Technology Showcase: Wireless Mics, The Digital Content Producer, 2005. [14] The Global Land One-Km Base Elevation Project (GLOBE), http://www.ngdc.noaa.gov/mgg/topo/globe.html. [15] The ITS Irregular Terrain Model Algorithm, NTIA, Department of Commerce. [16] OET BULLETIN No. 69, Longley-Rice Methodology for Evalu- ating TV Coverage and Interference. February 2004. [17] FCC press release, Evaluation of the Performance of Prototype TV-Band White Space Devices. November 2008. [18] P. Bahl, R. Chandra, T. Moscibroda, R. Murty, and M. Welsh. White Space Networking with Wi-Fi like Connectivity. In SIGCOMM, 2009. [19] R. E. D. Borth and B. Oberlie. Considerations for Successful Cognitive Radio Systems in US TV White Space. In DySpan, 2008. [20] J. Broch, D. A. Maltz, D. B. Johnson, Y.-C. Hu, and J. Jetcheva. A Performance Comparison of Multi-Hop Wireless Ad Hoc Network Routing Protocols (1998). In MOBICOM,1998. [21] G. Buchwald, S. Kuffner, M. Brown, and E. C. L. Ecklund. The Design and Operation of the IEEE 802.22.1 Disabling Beacon for the Protection of TV Whitespace Incumbents. In DySpan, 2008. [22] Y.-H. Chu, S. G. Rao, and H. Zhang. A Case for End System Multicast. In SIGMETRICS,2000. [23] S. Deb, V. Srinivasan, and R. Maheshwari. Dynamic Spectrum Access in DTV Whitespaces: Design Rules, Architecture and Al- gorithms. In MobiCom, 2009. [24] FCC Media Bureau. TVQ TV Database,
[25] S. Gaonkar, J. Li, R. R. Choudhury, L. Cox, and A. Schmidt. Micro-Blog: sharing and querying content through mobile phones and social participation. In MOBISYS,2008. [26] A. Goldsmith. Design and Performance of High-Speed Communication Systems over Time-Varying Radio Channels. PhD Thesis, University of California at Berkeley, 1994. [27] D. Gurney, G. Buchwald, L. Ecklund, S. Kuffner, and J. Grosspi- etsch. Geo-location Database Techniques for Incumbent Protec- tion in the TV White Space. In DySpan, 2008. [28] D. Katabi. The Use of IP Anycast for Building Efficient Multicast. In Global Internet, 1999. [29] H. Kim and K. G. Shin. In-band Spectrum Sensing in Cognitive Radio Networks: Energy Detection or Feature Detection? In MOBICOM,2008. [30] M. Mishra and A. Sahai. How much White Space is there? Technical report, EECS Technical Report, EECS-2009-3, January 2009. [31] S. Narlanka, R. Chandra, P. Bahl, and I. Ferrell. A Hardware Platform for Utilizing the TV Bands with a Wi-Fi Radio. In IEEE LANM AN,June 2007. [32] V. Otsason, A. Varshavsky, A. LaMarca, and E. de Lara. Accurate gsm indoor localization. In UBICOMP, 2005. [33] S. Ratnasamy, A. Ermolinskiy, and S. Shenker. Revisiting IP multicast. In SIGCOMM, 2006. [34] A. Sahai and D. Cabric. Spectrum sensing: fundamental limits and practical challenges. In DySPAN, 2005. [35] S. J. Shellhammer, A. K. Sadek, and W. Zhang. Technical Chal- lenges for Cognitive Radio in the TV White Space Spectrum. In Information Theory and Applications Workshop, 2009. [36] T. Shu and M. Krunz. Throughput-efficient sequential channel sensing and probing in cognitive radio networks under sensing errors. In MOBICOM, 2009. [37] R. Tandra and A. Sahai. SNR Walls for Signal Detection. IEEE Journal on Selected Topics in Signal Processing, February 2008. [38] W. Yu-chun, W. Haiguang, and P. Zhang. Protection of Wireless Microphones in IEEE 802.22 Cognitive Radio Network. In ICC Workshops, 2009. [39] Y. Yuan, P. Bahl, R. Chandra, P. A. Chou, I. Ferrell, T. Moscibroda, S. Narlanka, and Y. Wu. KNOWS: Kognitiv Networking Over White Spaces. In DySPAN, 2007.
threshold of -114 dBm. Furthermore, it is important that any practical solution does not require changes to the wireless mic system itself. This is a politically charged issue [5] since
Appendix
In this section we describe in further detail the design and implementation of the MicSentry. The primary purpose of such a device is to improve the liklihood of accurately sensing wireless microphones. This device is not necessary for the correct and efficient operation of SenseLess. Instead, this is an optimization that can help improve the liklihood of detecting wireless microphones. This is ap- plicable only when wireless microphones are not manually inserted into the database.
9.1 Detecting Wireless Microphones

The key challenge in sensing wireless microphones is that they must be accurately detected down to a
users usually spend a few tens of thousands of dollars [6] for planning wireless microphone frequencies. Hence, we want to minimize changes to existing systems. Our solution is beacon-based, i.e., we use a separate, cheap device, which we call the MicSentry, that is placed close to the wireless mic and hence, can detect the mic at much higher thresholds than -114 dBm. The MicSentry signals the presence of the mic to nearby WSDs either through the geo-location service (using an out-of-band 3G connection) or directly by transmitting packets. Using a separate device simplifies the mic detection problem, since a WSD now needs to detect a custom packet sent by a MicSentry rather than rely on mics that do not follow any known standard. This because a packet is a defined signature as opposed a typical mic signal that lacks a signature. The concept of signaling the presence of mics using special beacon signals is not particularly surprising, and has in fact recently been proposed [38, 4, 21]. What may be more surprising, however, is the fact that the question of how to beacon the presence of a mic is in fact a highly non-trivial problem, and none of the proposed solutions are in fact practically efficient. We outline these challenges in the subsequent section, and derive general design principles that govern the design of our own, practical beaconing solution: the MicSentry. 9.1.1 Beaconing Challenges There are many ways a beaconing device could transmit mic information, each with its own distinct set of challenges. In-band beaconing: Beacons signals are transmitted within the same TV-channel that is occupied by the wireless mic. There are various problems with in-band sensing. First, it must be ensured that the beaconing signal does not interfere with a second, more distant wireless mic that is concurrently using another part of the same TV-channel. This is a critical challenge since in effect, it implies that either every beaconer device must sense for other beaconing devices (notice that hidden-terminal problems become an issue in this case, too), or a separate portion of the spectrum of each TVchannel must be reserved for beaconing devices, which would require existing mic users to reassign their frequency assignment, so as not to use the reserved spectrum [5]. Secondly, as explained earlier, in-band sensing requires silent periods at all white-space nodes to reliable detection, which impacts the throughput of the network [36]. And finally, the MicSentry will have to transmit at significantly high power to signal reliably to the BS. Transmissions at such high power using commodity transmitters will have leak- age and can
interfere with mics on adjacent channels. Out-of-band beaconing in UHF spectrum: Out-of-
band sensing means the beacon to protect a mic is transmitted on a free UHF TV-channel, that is not the channel used by the mic. This can potentially eliminate silent periods, but it has drawbacks as well. First, in some cases, another channel may not be free, in which case a wireless mic cannot be protected. Second, outof-band beaconing implies out-of band sensing at clients, which requires an RF frontend dedicated to sensing (see Section 2). A third reason is particularly important for practical whitespace network. Out-ofband beaconing in UHF bands is fundamentally inefficient with regard to beacon detection/sensing at clients. In particular, the following theorem holds: In an out-of-band beaconing solution, any WSD must scan all possible TV-channels in order to guarantee it does not interfere with a mic. That is, a WSD can never simply scan one or a few TV-channels and be sure that it is allowed to transmit; it must always scan all channels before transmitting. To prove why the above claim is true, assume for con- tradiction that a white space device V wants to transmit on some channel CW S , and that it is sufficient for V to sense for beaconing signals on some subset C of channels. That is, we assume that V can tell for sure that there is no mic on CW S after checking that there are no beacons on any channel in C. For any such scenario, we can construct a worst-case in which V ends up interfering with a mic M on CW S as follows. Let V and M (and along with M its beaconing device) such that V is within M s protection range. M transmits on channel CW S . Further, let there be |C| other mics M1 , . . . , M|C| , each of which is located such that (i) Mi is within M s protection range, but (ii) V is outside the range of Mi s beaconing device. Notice that such a positioning is always (and easily) possible given practically possible sensing thresholds. In such a scenario, the beaconing device pro- tecting mic M is not allowed to transmit on any channel in C (because otherwise, it could interfere with a mic), but V does not have any knowledge about the existence of mics M1 , . . . M|C| and thus expects the beaconing device at M to beacon on one of the channels in C. Interference would occur, which proves the theorem. The implications of these findings are important. One could propose intelligent schemes in which beacon signals are transmitted on certain TV-channels, such that they are likely to be found much more easily by client devices. However, the above proof demonstrates that to protect against the worst-case, every white space device would still need to scan all channels. And in white space networks, the worst case matters since we must ensure non-interference to primaries. Design Questions: In view of the above, five design goals for a beaconing design can be distilled: (i) Beaconing signals must not interfere with other mics.
(ii) Beaconing devices must co-exist with other white
space devices. Ideally, no silence period is necessary. (iii) The solution should allow for low scanning/detection cost at white space devices (both in terms of scanning time and hardware cost). (iv) Detection must be guaranteed in spite of packet collisions (e.g. hidden terminals). (v) It should help increase the required detection threshold, in particular when signaling mics to BSs. (vi) Minimal changes to existing mic system setups. Prior proposals for beaconing-based mic detection do not achieve these goals.15 For example, approaches similar to the one explored by IEEE 802.22.1 [38] and others [4, 21] fails to achieve most of the above goals, e.g., they would require reassigning the mics frequency assignment (or, alternatively, would be at risk of interfering with other mics). Also, importantly, such a beaconer would have to transmit at significantly higher power to signal reliably to the BS (up to 250mW Tx power on the same channel as the mic [38]). Transmissions at such high power using commodity transmitters will have leak- age and can cause interference to mics on the same or adjacent channels. Note, even if a beaconing functionality is inducted into future mics, they will have to still follow the above listed design goals we propose. In the next section, we describe the design of MicSen- try a complete and practical beaconing-based solution for signaling mics to BSs, which achieves all the above properties. 9.1.2 The MicSentry The MicSentry has two main functions: detecting a mic, and signalling to a nearby WSD. For the first, we require the MicSentry to be placed close to the wireless mic receiver.16 Since the mic receiver is nearly always plugged to the wall power outlet, we expect it to be easier to place the MicSentry nearby. Furthermore, since the mic receiver is capable of decoding the mics transmission, a nearby MicSentry will be able to detect the signal at the FCC specified threshold of -114 dBm. In other words, MicSentry does not require lowthreshold sensing capa- bility. The challenging part is signalling to a nearby WSD. As discussed, both out-of-band and in-band signalling in UHF have significant, and in many aspects fundamental challenges. Instead, we achieve all design goals by using VHF white spaces for signalling the presence of the mic.
15 Furthermore, none of these proposals are fleshed out in detail. Hence specific analysis of such proposals is impossible. 16 Our current implementation of MicSentry also provides an extra layer of protection: once MicSentry detects the presence of a mic, it updates the geo-location service (see Section 4.5) with its location using a 3G connection. This provides mics protection via the geolocation service without the overhead of manually updating the service. But this obtaining such a 3G connection may not always be
feasible.
The VHF bands are in even lower frequencies than UHF (see Section 2.1), and have much better propagation characteristics. Note, we are not proposing to modify existing microphones to transmit over VHF frequencies. Instead, it is the MicSentry that transmits beacon packets over the VHF frequencies to signal the presence of a mic. There- fore, the MicSentrys signal will undergo less attenuation than that of the mic. Furthermore, since most mics only operate over UHF [13], many critical challenges are circumvented. So, these VHF channels, if available, are a very good medium for the MicSentrys signals. We evaluate the propagation characteristics and availability of VHF channels in Section 9.1.4. 9.1.3 MicSentry Design The VHF beacon must indicate which part of the UHF spectrum a mic is operating on. Hence, our design requires the BS to be able to decode the MicSentrys packets in order to detect a mic. The main facets of the Mic- Sentry design are: Frequency: A MicSentry uses the lowest available VHF channel available to transmit beacons (packets) signalling the presence of a mic. Each beacon contains the mics UHF channel number. Since there are only 30 UHF channels, this corresponds to 5 bits of information. WSDs scan all VHF channels (there are only 9, instead of 30 in UHF) to learn of nearby mics. Transmit Power and Decoding Threshold: Since the MicSentry can be classified as an auxiliary part 74 device, we set its transmit power to less the maximum per- mitted transmit power i.e. we set it to 40mW, the same transmit power we assume for a wireless mic.Also, since the MicSentry transmits over VHF frequencies, as a result of better propagation, the decoding threshold for its packets need to be calculated. This is achieved by the following conservative calculation. Using free space propagation, which is known to be very conservative, we first compute the distance at which the signal attenuation of a wireless microphone transmitting at 40mW over 512 MHz (the lowest UHF channel) will be -114 dBm. We then compute the attenuation for the same distance (since the MicSentry is colocated with the wireless mic) when the MicSentry transmits over the highest VHF channel (216 MHz). We find the threshold for decoding for the MicSentry is set to -106 dBm over VHF. Therefore, if a WSD is able to decode the MicSentrys packets down to - 106 dBm, it will ensure a -114 dBm protection of the wireless microphone. Bandwidth and Modulation: To decode packets at low SNR we transmit beacons using a narrow band-
width. Ideally, we would like to transmit these packets at the narrowest possible bandwidth, since decreasing the bandwidth also decreases the noise floor of the receiver
(Section 2). However, restricted by our current prototype hardware we are able to only achieve 200 kHz of the bandwidth. We use BPSK modulation for transmitting packets. Using these settings, we are able to achieve the low decodability threshold of -106 dBm. Co-Existence with Data Transmissions: A MicSentrys transmissions in the VHF band may face interference from other white space data transmissions in the VHF band. As discussed previously, possible solutions include blocking off an entire channel for the MicSentrys use, or silent periods, but neither option is desirable. The problem is that if we allow white space devices to continue operating as usual, they may not carrier sense the MicSentrys narrowband transmission. This is because the sensed energy is averaged across the FFT bins spanning the entire channel. To address this challenge, we constrain a MicSentry to only transmit at the center of a 6 MHz TV channel. We then add a separate component on the receive path of a white space device to enable it to detect the presence of the narrowband MicSentry. After performing the FFT of an incoming signal, in addition to carrier sensing across the entire channel, the receiver measures the power only in the FFT bins around the center of the 6 MHz channel to estimate whether there is a MicSentry that is actively transmitting. If this is the case, the white space device employs the regular backoff mechanism used by its carrier sensing mechanism. Note, this is only performed if the medium is deemed to be idle when carrier sensing across the entire channel. At the MicSentry, we always perform carrier sensing in the 200 kHz narrowband at the center of the VHF channel which lowers its noise floor. Hence, the MicSentry is able to carrier sense the presence of other MicSentry devices as well as white space devices. Since a MicSentry co-exists with other white space devices operating in VHF, we must ensure that inspite of collisions (e.g. hidden terminals), there is a non-zero probability of the BS decoding the MicSentrys packets. This is to ensure that a mic will not go undetected. One way to achieve this is to give the MicSentry prioritized access to the medium. This can be achieved by reducing the wait time between two successive transmissions of the MicSentry (similar to slot time in 802.11 networks) . Since a MicSentry is expected to use very small packets, each time the MicSentry beacons, it sends a burst of a few packets one after the other thereby improving the likelihood of detection at the BS. Since a MicSentry beacons packets infrequently (100ms in our implementations), this approach does not saturate the medium. We note however, this approach limits the throughput of other white space devices contending with the MicSen- try over VHF white spaces (but not UHF). But this tradeoff may be necessary to ensure
non-interference to mics.
Detection Probability
We implement this feature and demonstrate its efficacy in Section 9.1.4. Another way to achieve the same goal is to use a RTS/CTS-like mechanism whereby each time the MicSentry attempts to transmit a beacon packet, it first broadcasts an RTS packet. 9.1.4 Commercial Viability of MicSentry The concept of a separate device to protect the primary users is not new. The FCC distributed the converter box for backward compatibility after the DTV transition. Similarly, mic manufacturers such as Shure [4] and Mo- torola [21] are proposing the use of a separate device to protect mics as well. However, these proposals use in- band sensing over UHF which is inefficient for the reasons discussed previously. One could argue implement- ing sensing at the white space clients may be cheaper than handing out MicSentrys to the mic users. How- ever, we note there are far fewer wireless mics (on the order of a few hundred thousand in the US including ille- gal mics) compared to tens of millions of potential white space clients if it were to become as successful as Wi-Fi. Furthermore, many mics are co-located at a venue, such as an auditorium, in which case multiple MicSentrys may not be required. For these reasons, the trade-off of reducing significant complexity from clients (no sensing required) at the cost of a low-complexity (no low-threshold sensing required) device for mics seems preferable.
Count
1 8
w
C i t i e s 1 5 S u b u r b s 1 2 T o w n s 9 6 3 0 0 1 2 3 4 5 6 7 N u m b e r o f c h a n n e l s a v
P r i o r i t i z a t i o n w / o P r i o r i t i z a t i o n 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Cro sstraff ic Pac ket Inte rval (ms )
Evaluating the MicSentry

Signalling over VHF white spaces has several benefits (Section 9.1.3). Hence, we first evaluate the feasibility of signalling over VHF. We then evaluate our prototype MicSentry decoding packets at the calculated -106 dBm. Feasibility of Transmitting over VHF First, via empirical outdoor measurements, we demonstrate the gain achieved by transmitting over VHF, as opposed to UHF. Our setup consisted of a node located on a building rooftop, transmitting at 4W over the UHF and VHF white spaces. Using a spectrum analyzer with a UHF/VHF antenna we measured the RSSI of the transmitter at various points within a 2 KM radius. For each location, we computed the difference in RSSI between VHF and UHF. The results illustrated in Figure 16, show a median gain of 14 dB over VHF. Also, notice most measurements between the upper and lower quartile lie between 12 - 16 dB, which is close to the 12 dB increase in threshold calculated earlier in Section 9.1.3. Hence, by enabling the MicSentry to transmit over VHF frequencies, we can raise the microphone detection threshold at a WSD.
CDF
a i l a b l e
(a) Predicted VHF Availability across 105 locations
1 0.75 0 . 5 0.25 0
2 4 6 8 1 0 1 2 1 4 1 6 1 8 G a i n i n R S S I ( d B )
-95 -85
-90 RS SI (db m)
separate USRP2 node to function as a receiver. The MicSentry
Figure 18: Probability of

detecting a mic as a function of wideband cross traffic rate, using a MicSentry in a hidden terminal situation
Figure 17: Decodability of

beacon packets at low SNR and varying bandwidths using the MicSentry prototype.
(b) Mea sure d Gain in RSS I whe n trans mitti ng over VH F vs UH F
But currently are there VHF white spaces in the US? After all, there are fewer VHF channels (9) for white spaces when compared to the number of UHF channels (30). We use the geo-location service to predict the availability of VHF white spaces across the 105 locations in the US. As seen in Figure 16(a), over 93% of the locations have at least one free VHF channel available. A MicSentry only requires one free VHF channel since it coexists with other white space and MicSentry devices using the same channel. Hence, we conclude VHF white spaces are available and it is feasible to use them for the MicSentry to signal to a WSD. Decoding beacon packets at Low SNR To evaluate decoding beacon packets at -106 dBm, we use our prototype MicSentry. It first detects a mic and then broadcasts this information in 16 byte beacon packets at regular intervals of 100ms. We use a
Figure 16: Case

for using VHF with the MicSentry
Fraction of Packets Received
100 80 60 40 200 KHz 20 800 KHz 0 1.6 MHz -115 -105 -110 -100
400
transmits a fixed number of beacon packets and at the receiver we measure the reception rate. Of the 16 bytes, 1 byte is used to convey the UHF channel used by the mic. The remaining bytes are added as FEC to improve reception rates at low SNR. We vary (i) the attenuation between the MicSentry and the receiver using the variable attenuator and (ii) the bandwidth of the narrowband beacon packet. For each run of the experiment we set the receivers sampling bandwidth to be equal to the bandwidth of the narrowband beacon packet. The results shown in Figure 17 have two highlights. First, at narrower widths we are able to decode more packets at lower RSSI. This is because, as described in Section 2.2, by lowering the sampling bandwidth at the receiver, we lower its the noise floor and improve its ability to decode packets at low SNR. Second, at a 200 kHz width we are able to successfully decode over 80% of the beacon packets at the prescribed threshold of -106 dBm. Therefore, by imposing a fixed bandwidth across all Mic- Sentry devices (200 kHz), a BS with a
radio whose sampling bandwidth is set to 200 kHz can reliably decode beacon packets at low SNR over VHF which protects the mic at the required -114 dBm over UHF. D e t e c t i o n L a t e n c y We measure the time taken for a BS to detect a mic by decoding a MicSentrys packets in a hidden terminal scenario. As described in Section 9.1.3, this is the worst case scenario for the MicSentry. In such situations the detection latency primarily depends on the probability of the MicSentrys packets being decoded successfully at the BS. We study this via the following simulation in Qual- net [7]. Based on the 802.11 PHY and MAC, we implemented the MicSentry including support for 200 kHz narrowband transmissions in Qualnet. The simulation setup consisted of a hidden terminal with a MicSentry (narrowband
200 kHz tx) at one end, a white space device (wideband 6 MHz tx) at the opposite end, and a BS at the center. The white space device and the MicSentry are out of sensing range of each other but in communication range of the BS at the center. All nodes use an
802.11 MAC and PHY17 . The white space device transmits a controlled rate of cross-traffic (128 byte packets) to the BS. As described in Section 9.1.3, the MicSentry is prioritized by reducing its slot time to 2s and it broadcasts a burst of 4 packets (each of 16 bytes with FEC) every 100ms. We vary the cross-traffic rate by changing its inter packet latency at the white space device. We fix the total number of beacons transmitted by the MicSentry and measure the reception rate at the BS to estimate the detection probability. The results in Figure 18 show that in the worst case, at a very high rate of cross-traffic, the detection probability with prioritization is 0.23 which, at a 100 ms beacon rate is an expected detection latency of 434 ms. Hence, prioritizing the MicSentrys traffic significantly improves the likelihood of detecting its beacons even in hidden terminal scenarios.
9.2 Discussion
The MicSentry, by virtue of using packet decoding (as opposed to energy or feature detection) and a higher threshold over VHF has the potential to improve microphone detection. The design and implementation of this device ensures that is operates smoothly with geolocation service. In other words, using information from the MicSentry, obtained either from the MicSentry directly or via WSDs that detect the beacon packets from the MicSentry, the SenseLess system can improve its management of channels. For example, if the MicSentrys packet format were to include the narrow band over which a nearby mic is operating, the SenseLess database can instruct nearby WSDs to only avoid those particular frequencies instead of blocking out the entire channel. Hence, adding a device such as the MicSentry has the potential to optimize the functioning of the SenseLess even further.
17 We use 802.11 because there does not yet exist a standard PHY/MAC for the white spaces

Senseless TR

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Senseless TR

Uploaded by

Copyright:

Available Formats

SenseLess: A Database Driven White Spaces Network

2 Backg round & Motivation

2.1 White Spaces

The FCC ruling in its entirety is a vast set of specific regulations.

is even considering a threshold of -126 dBm in the UK.

3.1 Determining White Spaces Availability

Figure 1: A 1500 mile path over which we measu red UHF

Egli Free Space

when measuring channel power values.

Fraction of white spaces lost (%)

Figure 2: Comparison with ground truth for diffe rent

This is a limitation imposed by our spectrum analyzer hardware

3.2 Need for Location

Cities Suburbs Towns 0 20 40 60 80 100

0 20 40 60 80 100 % UHF Channels Lost

% UHF Channels Lost

White Space Loss(%)

1 Cities 0.75 Suburbs 0.5 Towns 0.25 0

1 0.75 0.5 0.25 0

100 80 60 40 20 0 0 0.5 1 1.5 2 2.5 3 3.5 4 Location error (miles)

% UHF Channels Lost

Figure 3: The impact of using BS location as a proxy for

Figure 4: The impact of location errors on the loss of white

3.3 Handling Mobility

Figure 5: Impact of mobility on the protection region

(a) SenseLess architecture overview

SenseLess Engine gation Models ... Longley Rice

4 SenseLess: System Design

4.1 System Overview

We intend to open up the service for public use by NSDI 2010.

Table 1: API exposed by the SenseLess service.

4.2 SenseLess Service

4.3 Back-end Store

4.4 SenseLess Engine

For TV towers with 1MW Tx power, an antenna height of

1 0.75 0.5 0.25 0

Interval Between Changes (days)

Figure 7: Changes to entries in the FCC TV transmitter

4.5 Detecting Microphones

4.6 System in Operation

been studied in prior work [18, 23].

very high frequencies.13

Additional channels gained 0 1 2

Table 2: Ground truth measu rements of additional white

Cities Suburbs Towns

Figure 8: Screenshot of a web based front-end GUI for

Number of additional channels gained

Figure 9: Predicted gain in white Spaces from variable

6.1 Variable Thresholds

Elevation Source Adaptive SRTM only GLOBE only

Category Cities Suburbs Towns

Table 3: Impact of various elevation sources on the false

Table 4: The minimum number of channels needed to

1 0.8 0.6 0.4 0.2 0

Figure 10: Size of compressed beacons as computed across

6.3 Reducing Beacon Sizes

We elide a map of the driving path due to a lack of space

6.2 Countering Holes in Elevation Data

6.4 System Scalability

Time to fetch white spaces availability (s)

1 0.8 0.6 0.4 0.2 0

Cities Suburbs Towns