You are on page 1of 7

l l l l l

������ �� ��� �������


� ������ �� ������� �������������� ��� ������������

������ ����������
��� � ������ �������� ��� ������� � �������

��������� l ������ ����

�������� ���� ������ l ��� ������ ����� l ��� � ��� ����� ������ l ���������� ������� ��������� ������� ��� ��� l �������������������������������
Technical White Paper Series
Duplex Conflicts

Company

Apparent Networks is a leading innovator of network intelligence software. Apparent Networks’s technology, AppareNet, a network
intelligence system, operates non-intrusively on live networks, to and from any location worldwide. Without requiring specialized
hardware or remote agents, AppareNet views the network from the application’s perspective. In doing so, AppareNet rapidly
identifies the locations and causes of network bottlenecks anywhere in the world so that companies can boost the performance
of, and gain more value from, the network infrastructure they already have. Apparent Networks improves its customers’ businesses
by helping organizations reduce operational costs, increase IP availability, and protect revenues.

Contact Information

Canadian Head Office

400 - 321 Water Street


Vancouver, BC
Canada V6B 1B8

Sales: 1.800.508.5233
Support: 1.800.664.4401
Telephone: 604.433.2333
Fax: 604.433.2311

http://www.apparentnetworks.com
marketing@apparentnetworks.com

This report in whole or in part may not be duplicated, reproduced, stored or retransmitted without prior written permission of Apparent Networks. All opinions and
estimates herein constitute our judgment as of this date and are subject to change without notice. Any product names mentioned herein may be trademarks and/or
registered trademarks of their respective companies.
Technical White Paper Series
Duplex Conflicts

Technical Series Duplex Conflict/Mismatch

This is the second in a technical series of white papers from Apparent A duplex conflict or mismatch is the mixing of transmission rules
Networks examining the Perils of the Network. This technical series between two network interfaces that are connected together. Each
explains network idiosyncrasies and degradations and how AppareNet interface implements a different rule set, resulting in simultaneous
is capable of identifying network problems. attempts at transmission causing lost packets.

The first paper in this series, The Apparent Network, introduced the
concept of the AppareNet network as a complete end-to-end view of Impact on Application Performance
a network. This paper discusses one of the most pervasive problems
on networks today, duplex conflicts. Some applications (e.g. ping, telnet) are rarely affected by duplex
problems. Links that handle only one socket at a time usually are
not impeded. Hence, the scale and performance impact of the
problem is a relatively minor issue for a network link that is used
only by non-intensive or one-way data exchange.
Executive Summary
A recent study concluded that duplex conflicts (otherwise known as However, when multiple applications (sockets) attempt to transmit
duplex mismatches) cause 75% of all Internet/2 problems.* Most simultaneously on a link with a duplex mismatch, packets are lost
Network Engineers have encountered such conflicts, and understand and re-transmissions are forced; applications begin to slow down
how elusive they can be. These conflicts often go undiscovered by and sometimes fail. This is a common scenario for application or FTP
existing tools and network management technologies, leaving users servers, and performance degradations of 50-80% from expected
experiencing painfully slow and erratic network performance. Some levels are quite common.
networking professionals have gone as far as to call the problem an
epidemic. One symptom of a potential duplex problem in an IP network is FTP
performance fade. When an FTP download is started, the initial
This paper defines the problem, along with some common performance and data transfer rate is acceptable, but then, as
application performance symptoms of duplex problems. A brief collisions and packet loss start to occur due to a duplex conflict, the
history of Ethernet technology provides a context for how and why TCP sliding window is progressively decreased until the performance
duplex conflicts have become a prevalent performance inhibitor in slows to a crawl.
today’s network infrastructures. The nature of conflicts on hubs and
switches, as well as an exploration of inconsistent "auto-negotiate" Before AppareNet, there simply were no tools that reported duplex
implementations, will also be outlined. conflicts. It was nearly impossible to understand when or where
conflicts and link degradations were occurring.
AppareNet offers unique capabilities for rapidly measuring network
capacity, often identifying the duplex mode, and can identify and A brief look at the history of Ethernet reveals how and why duplex
pinpoint duplex conflicts in an IP network. Once the problem source conflicts occur as often as they do.
is isolated, a simple reconfiguration of the network equipment is
typically all that is required to achieve significant and noticeable
improvements in IP network performance. AppareNet allows
organizations to maximize the performance and ROI of IP networking
infrastructures.

* Internet2, Fall 2001, End-to-end Performance Initiative

Page 1
Technical White Paper Series
Duplex Conflicts

Brief History of Ethernet Hub

Ethernet originally was either 10Base-2 or 10Base-5. Both flavors ** Hub Internal
Connection
** Hub Internal
Connection
** Hub Internal
Connection

used transceivers connected via coaxial cable, carrying only one * Xceiver

transmission at a time. Network Interface Card (NIC) transceivers


joined the transmit (Tx) and receive (Rx) wires together. Tx Rx

Port Port Port Port

Tee Connector Tee Connector Tee Connector Tee Connector


Xceiver

** Coaxial Cable ** Coaxial Cable ** Coaxial Cable

Rx Tx
* Xceiver

NIC a NIC b NIC c NIC d

* Tx and Rx joined here in the Hub


** Only one transmitting station allowed at one time;
internal function similar to a coaxial cable
Tx Rx

In 10Base-T networks, hubs typically interconnect NICs using cables


NIC a NIC b NIC c NIC d
that have Tx and Rx transmission pairs. What is not clearly evident
* Tx and Rx joined here in the NIC
is the fact that within the hub, the Tx and Rx wires are connected
** Only one transmitting station allowed at one time
together. Functionally and electronically, 10Base-T hubs work exactly
the same as 10Base-2 and 10Base-5 Ethernets. In other words, they
operate in half-duplex mode.
When networks were connected in this fashion (as illustrated above),
the transmission protocol used was CSMA/CD (Carrier Sense Multiple
Access/Collision Detect), also known as half-duplex (HDx). Simply Switch
stated, a station does not transmit if it detects any other stations Switch Fabric

transmitting. While it is transmitting, it does not listen for other


stations. It does listen for collisions (multiple stations transmitting at
once), subsequently retransmitting at a later time.
Tx Rx

Reliability and connectivity problems led to the next generation of Port Port Port Port

Ethernet employing 10Base-T.


Xceiver

Rx Tx

NIC a NIC b NIC c NIC d

In switches, the Tx and Rx wires are independent of each other,


allowing simultaneous multi-direction transmissions. For the first
time, CSMA/CD rules became optional. Full-duplex connections are
supported, allowing any NIC to transmit at any time that it had data
available for transmission.

And this is where the duplex conflicts began.

Page 2
Technical White Paper Series
Duplex Conflicts

Conflict on Hubs Conflict on Switches

Once NICs started to support full-duplex (FDx), hub-based systems Switches require setting of duplex mode on a per-port basis. The port
were affected by any NICs that were set to full-duplex. setting must match the settings of the device attached to the port.

������ �������� �� ���� �� �������� � ���� �������� ��� �� ����������� ������ �������� �� �������� �� �������� � ���� ��������� ������ �������� ������� ���� ��
������� � ������� � ��� ������������� ������ ����

Duplex Conflict on Hubs Conflict - FDx not allowed on hub ������ �������� �� �������� �������� � ���� ����� �� � ���� ���� �� ��� ��� ����
Diagram B ������� �

Recall that a hub can only support one transmitting station at a time. In Duplex Conflict on Switches - Diagram A, traffic will flow smoothly
When Station B (in Diagram B) is set to full duplex, it may attempt to between Station A and Station B because each station’s duplex setting
transmit while receiving other traffic. This corrupts any frames that matches that of the corresponding switch port.
are passing through the hub, resulting in packet loss.
In Diagram B, traffic will flow correctly between Station A and its
switch port. However, if Station A sends data to Station B, the duplex
conflict between Station B and the switch can cause packet collisions.
Station B does not detect collisions. Station B’s full-duplex setting will
cause it to transmit regardless of the state of its receive line. As Station
B’s switch port is set to half-duplex, restricting the port from receiving
while it is transmitting, Station B’s packets will be lost.

Page 3
Technical White Paper Series
Duplex Conflicts

Auto-Negotiate Only by explicitly setting both sides of the link to the same duplex
mode would the link work at optimum performance.
Supposedly the solution to the duplex conflict, auto-negotiation has
actually made the problem even more insidious. Typically duplex To deal with this problem, many switches have received microcode
auto-negotiation techniques for NICs, switches and routers are either updates that change the behavior of their auto-negotiate mode. This
obscure or incorrectly implemented. revised mode of auto-negotiation does not appear to guess what the
other end of the link is set to. Rather, the port operates in a mode that
Most of us have seen auto-negotiate settings on NICs, switches and is neither half-duplex nor full-duplex. Although performance is
routers. Logic suggests that using auto-negotiate settings either at degraded, the possibility of a duplex conflict and subsequent packet
the NIC or at a port will ensure proper negotiation of the duplex loss is less likely.
setting. The sensible outcome would be if the interfaces settled on
the most favorable duplex setting. Apparent Networks has set up test cases where a station is set to full-
duplex mode, and the switch is set to auto-negotiate. We do not see a
conflict between half and full-duplex, but AppareNet found that the
The problem is that there is no clear standard that defines auto- capacity of the link was 35% below expected; yet higher than would
negotiation. In a practical sense, we have seen many implementations be possible on a half-duplex link. By switching from auto-negotiate
and interpretations of this so-called feature. to full-duplex, the full capacity of the link was once again realized.
This example of auto-negotiation sacrificed performance to address
the inability to accurately negotiate duplex modes.
Commonly, auto-negotiate means that, when the link is established
(NIC is connected to a port), negotiations will occur to ensure that
both sides of a connection are set to the right speed and duplex Measuring Path Bandwidth
mode. The speed is often correct, however, we often find that duplex
negotiations fail. Unlike most other technologies that can measure traffic throughput
or "what the network is doing", AppareNet is unique in its ability
to measure "what the network can do" from API to API along any
IP test path. AppareNet can also distinguish whether an end-to-end
path is working at half or full-duplex.

The diagram below is an example where AppareNet has measured the


last hop of an end-to-end test path as half-duplex Ethernet.

������ �������� �������������� ��� ���� � �������� ����� �� ������ ��������������

The illustration above is an example of this behavior on a common


switch device that Apparent Networks tested with AppareNet. We
set the 100 Mbps switch ports to auto-negotiate mode, and tested
with NICs from three different vendors. We set each card to auto-
negotiate, half-duplex and full-duplex. In each case, a duplex conflict
occurred. Furthermore, the switch often reported the negotiated The reported bandwidth is slightly less then the expected 9.75 Mbps
duplex mode incorrectly. For example, if the NIC was set to full-duplex, (Ethernet 10 Mbps half-duplex less IP overhead). The slight variations
the switch would report that it had negotiated full-duplex, however in bandwidth measurement are dependent on clock speed in the NIC
AppareNet still identified that a duplex conflict was present. of the target device.

Page 4
Technical White Paper Series
Duplex Conflicts

Here is an example where AppareNet reports a full-duplex Ethernet Once resolved, AppareNet can run the test again to verify a clean
bandwidth: network path (i.e. Full-duplex 100 Mbps Ethernet).

Again, the measured capacity is slightly less then the theoretical


maximum of 19.5 Mbps (a two-way measure) for a 10 Mbps full-
duplex link.
Resolving Duplex Setting Problems

Apparent Networks encourages and recommends that most


enterprises with large networks should not use auto-negotiate.
Finding Duplex Setting Problems However, in practice it is difficult to enforce in dynamic networking
environments. A related challenge is that several brands of network
Existing tools and network management technologies have been adapters default to auto-negotiate at installation time, thus making
unable to adequately identify and control the pervasiveness of
conflicts more likely.
duplex problems in IP networks.

AppareNet is the perfect tool for taking control of duplex conflicts.


Regular testing of your IP network using AppareNet reveals
AppareNet’s sophisticated analysis of test packets, after they return to
conflicts and slowdowns on your network instantly. The test reports
the AppareNet sequencer, (the single-end test measurement probe)
pinpoint exactly where network technicians need to apply fixes to
allows the system to identify problem "signatures", and report
duplex settings on network equipment. AppareNet allows network
findings to the user with clear diagnostic information.
operators to identify and eliminate duplex problems, and maximize
the performance and ROI of existing infrastructure.
The following is an example of AppareNet’s identification of a duplex
conflict.

Emerging Ethernet Standards


AppareNet can uniquely identify which interface has which duplex
setting. "Half/full-duplex" indicates that the half-duplex setting
As a final note, 1 Gbps links work only in full-duplex mode and the 10
is on the interface closest to the sequencer. "Full/half-duplex"
Gbps standard disallows half-duplex operation. However, anticipate
indicates the converse.
that Gbps Ethernet’s jumbo frames will be responsible for a whole
new raft of issues, such as MTU misalignments.

AppareNet can also be leveraged to identify and isolate MTU


misalignments. This is a separate white paper topic from Apparent
Networks.

For further information on AppareNet, or to see it live, please contact


us at marketing@apparentnetworks.com or toll free at 1.800.508.5233
AppareNet has isolated the half/full-duplex problem at a particular or visit our website at www.apparentnetworks.com.
IP address. Hence, a technician can now quickly isolate the location
of fault, and apply the necessary adjustments to the settings on one
or both ends of the problem link.

Page 5

You might also like