Microsoft SQL Server 2008 R2 Technologies for SAP Solutions

Applies to NetWeaver 7.00 (SR3) and 7.10 based products Summary This white paper provides an overview of the key features of Microsoft SQL Server 2008 R2 and the key
features of SQL Server 2008 that are relevant to SAP NetWeaver 7.00 (SR3) and 7.10 based products.

Author Jrgen Thomas, Principal Program Manager Company Microsoft Corporation Author Bio Jrgen Thomas, Microsoft, Principal Program Manager Technical Reviewer Cameron Gardiner, Microsoft, Senior Program Manager

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 2 of 22

Disclaimer and Liability Notice

This document may discuss sample coding or other information that does not include SAP official interfaces and therefore is not supported by SAP. Changes made based on this information are not supported and can be overwritten during an upgrade. SAP will not be held liable for any damages caused by using or misusing the information, code or methods suggested in this document, and anyone using these methods does so at his/her own risk. SAP offers no guarantees and assumes no responsibility or liability of any type with respect to the content of this technical article or code sample, including any liability resulting from incompatibility between the content within this document and the materials and services offered by SAP. You agree that you will not hold, or seek to hold, SAP responsible or liable with respect to the content of this document. The information contained in this document represents the current view of Microsoft Corporation on the issues discussed as of the date of publication. Because Microsoft must respond to changing market conditions, the information presented herein should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information presented after the date of publication. This white paper is for informational purposes only. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED, OR STATUTORY, AS TO THE INFORMATION IN THIS DOCUMENT. Complying with all applicable copyright laws is the responsibility of the user. Without limiting the rights under copyright, no part of this document may be reproduced, stored in, or introduced into a retrieval system, or transmitted in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise), or for any purpose, without the express written permission of Microsoft Corporation. Microsoft may have patents, patent applications, trademarks, copyrights, or other intellectual property rights covering subject matter in this document. Except as expressly provided in any written license agreement from Microsoft, the furnishing of this document does not give you any license to these patents, trademarks, copyrights, or other intellectual property. Unless otherwise noted, the example companies, organizations, products, domain names, e-mail addresses, logos, people, places, and events depicted herein are fictitious, and no association with any real company, organization, product, domain name, e-mail address, logo, person, place, or event is intended or should be inferred. 2011 Microsoft Corporation. All rights reserved. Microsoft, the Microsoft logo, Azure, BizTalk, Excel, SharePoint, SQL Server, and Windows are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries. All other trademarks are property of their respective owners.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 3 of 22

Table of Contents
Executive Summary ................................................................................................................................................................................................ 5 Who Should Read this White Paper ........................................................................................................................................................... 5 Topics Covered in this White Paper ........................................................................................................................................................... 5 SQL Server 2008 R2 Release ............................................................................................................................................................................... 6 New Features in SQL Server 2008 R2......................................................................................................................................................... 6 UCS-2 Compression for Unicode Data Type ..................................................................................................................................... 6 Support for Machines with 256 Logical Processors ........................................................................................................................ 7 Support for Up to 15,000 Table Partitions .......................................................................................................................................... 7 SQL Server 2008 Release ..................................................................................................................................................................................... 8 Key Features in SQL Server 2008 ................................................................................................................................................................. 8 Features Actively Leveraged by SAP NetWeaver ............................................................................................................................. 9 Transparent Features of SQL Server 2008 .............................................................................................................................................. 14 Backup Compression ................................................................................................................................................................................. 14 Audit ................................................................................................................................................................................................................ 15 Transparent Data Encryption ................................................................................................................................................................. 16 Encrypt Connections using Secure Sockets Layer ......................................................................................................................... 16 Lock Escalation Granularity ..................................................................................................................................................................... 17 Minimal Logging Enhancements .......................................................................................................................................................... 17 Query Plan Freezing .................................................................................................................................................................................. 18 SQL Server Database Mirroring Enhancements Compression .............................................................................................. 19 SQL Server Database Mirroring Enhancements AutoRepair .................................................................................................. 19 Changed Autogrowth Behavior ............................................................................................................................................................ 20 Appendix .................................................................................................................................................................................................................. 21 SAP and SQL Resources ................................................................................................................................................................................ 21 SAP Notes ........................................................................................................................................................................................................... 21 Microsoft Blogs................................................................................................................................................................................................. 22

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 4 of 22

Executive Summary
Microsoft and SAP have been working together closely for more than 15 years to ensure that the Microsoft platform and SAP solutions work well together. SAP certification of Microsoft Windows Server 2008 and Microsoft SQL Server 2008 R2 is the technical stamp of approval of the joint work between Microsoft and SAP. Worldwide, the number of developers proficient in these technologies and the number of organizations adopting SQL Server for SAP continue to grow.

Who Should Read this White Paper

This white paper is intended for IT professionals and technical decision makers looking for a database solution to run their SAP application.

Topics Covered in this White Paper

This white paper discusses some of the feature enhancements to SQL Server 2008 and SQL Server 2008 R2 that solidify the position of SQL Server as an enterprise-ready platform for SAP. Key highlights of the Microsoft and SAP partnership include: SQL Server and SAP offer a reduced total cost of ownership (TCO) for database platforms through lower pricing, reduced administration overhead, built-in high availability, and scalability. Microsoft Windows and SQL Server run on powerful new Intel and AMD processors that (as of an April, 2011 benchmark) recorded at 73,970 SAPS and 13,550 benchmark users on a standard HP ProLiant BL680c i G7, 4 processor commodity server for the SAP ERP Sales and Distribution (SD) benchmark. SAP and Microsoft fully support SQL 2008 and SQL 2008 R2 in virtual environments. For more information on Windows Virtualization, see SAP note 1409608. SAP and Microsoft now support up to 15,000 table partitions for SQL Server 2008 SP2 and SQL Server 2008 R2 SP1.

This white paper is organized in two sections. The first section discusses enhancements to SQL Server 2008 R2 and the second section discusses enhancements to SQL Server 2008.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 5 of 22

SQL Server 2008 R2 Release

Microsoft released SQL Server 2008 R2 for SAP NetWeaver 7.0-based systems in September, 2010. This release offers two major feature enhancements to make SAP run efficiently on SQL Server. This section provides a detailed explanation of the new features in SQL Server 2008 R2 and what it means for a SAP installation. Note: SAP treats the SQL Server 2008 R2 release more as a service pack than as a separate release. As a result there will not be an explicit entry for SQL Server 2008 R2 in the SAP Product Availability Matrix (For more details, see SAP note 1076022). Instead, the SQL Server 2008 entries automatically include both SQL Server 2008 and SQL Server 2008 R2 support.

New Features in SQL Server 2008 R2

SQL Server 2008 R2 includes the following three major features: UCS-2 compression for Unicode data type Support for machines with 256 logical processors (an increase from 64 to 256) Support for up to 15,000 table partitions with SQL Server 2008 R2 SP1 and SQL Server 2008 SP2

UCS-2 Compression for Unicode Data Type

A new feature in SQL Server 2008 R2 is UCS-2 compression for Unicode data type. SQL Server uses UCS-2 (instead of UTF-8) to encode Unicode characters. Unicode compression is an improved algorithm to store Unicode characters when using row/page compression and not a new compression type. SQL Server 2008 R2 stores row or page compressed UCS-2 characters more efficiently than SQL Server 2008. This feature is called UCS-2 compression or simply Unicode compression. UCS-2 compression significantly reduces the disk space needed to store the Unicode characters. The amount of disk space that gets reduced by UCS-2 compression varies by installation type. Figure 1 illustrates compression ratios achieved on a 5 terabyte customer database for a SAP ERP Unicode system. Index fragmentation did not have any impact since the database was completely reorganized before the compression.
Figure 1. Compression Ratios Compression Type No compression Row compression on data (clustered index) using SQL Server 2008 Page compression on data (clustered index) using SQL Server 2008 Row compression on data (clustered index) using SQL Server 2008 R2 Row compression on data and index (clustered and non-clustered index) using SQL Server 2008 R2 Page compression on data and index (clustered and non-clustered index) using SQL Server 2008 R2 Percentage of Original 100 percent 83.35 percent 46.36 percent 64.28 percent 54.75 percent 24.88 percent

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 6 of 22

Note: For a detailed discussion on database compression for SAP, see the Using SQL Server Database Compression with SAP NetWeaver white paper at http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/50c5de8f-282a-2e10-cb8aa5859a0994a4?QuickLink=index&overridelayout=true.

Support for Machines with 256 Logical Processors

SQL Server 2008 R2 increased the number of physical and logical processors it supports. SQL Server 2008 R2 Enterprise edition x64 now supports up to 8 physical processors and 256 logical processors. SQL Server 2008 R2 Datacenter x64 supports more than 8 physical processors (as an option). Note: For more information about processor support, see http://blogs.msdn.com/b/saponsqlserver/archive/2011/04/18/how-many-logical-processors-does-sql-server-2008r2-enterprise-edition-support.aspx.

Support for Up to 15,000 Table Partitions

Customers running the SAP Business Warehouse (SAP BW) who have encountered the 1,000 partition limit with previous versions of SQL Server can take advantage of the new 15,000 partition limit with SQL Server 2008 SP2 and SQL Server 2008 R2 SP1. After installing the service back, you need to enable the feature by issuing the following Transact-SQL command against your specific database: exec sp_db_increased_partitions '<database-name>', 'ON' Running this command increases the schema version for the database. This has implications when upgrading from SQL Server 2008 SP2 to SQL Server 2008 R2 SP1. To complete the upgrade process, follow these steps: Detach the database Upgrade the instance to SQL Server 2008 R2 Apply Service Pack 1 for SQL Server 2008 R2 Attach the database

Note: SAP has published the correction instructions for SAP BW in SAP note 1494789, Enabling 1000+ partitions support on SQL Server. For more information, see http://blogs.msdn.com/b/saponsqlserver/archive/2011/03/18/increased-partition-support-of-sql-server-in-sapbw.aspx.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 7 of 22

SQL Server 2008 Release

Microsoft released SQL Server 2008 for SAP NetWeaver 7.0-based systems in January, 2009. This section introduces major new features in SQL Server 2008 that will benefit SAP customers. These features are divided into the following: Features that require adaptations to be completed with SAP coding Features that are transparent to SAP application stacks

To take advantage of these features, you need to run SAP applications based on NetWeaver 7.0 and meet the requirements specific to the basis support packages. For more information, see SAP note 1152240.

Key Features in SQL Server 2008

SQL Server 2008 provides the following features that are actively used by SAP coding: New Open Database Connectivity (ODBC) client interface Data compression Star join optimization for business intelligence (BI) only

The following features are helpful when running a SAP system: Backup compression Auditing Transparent data encryption Encrypt connections using Secure Sockets Layer (SSL) Lock escalation granularity Logging enhancements Resource Governor Query Plan freezing Database mirroring enhancements Changed Autogrowth algorithms

The features are transparent to SAP applications and need no further changes in SAP coding.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 8 of 22

Features Actively Leveraged by SAP NetWeaver

ODBC Client Interface With SQL Server 2008, SAP re-implemented the database interface and switched from OLE DB to ODBC. The main reasons for this switch are: ODBC is the industry standard for database interfacing and the most widely-used and optimized technology. The former OLE DB interface has a ten-year history starting with SQL Server 7.0. Since more features are offered by newer database releases and functionality moves from the interface into the database server, a significant amount of code becomes obsolete but still has to be maintained for compatibility reasons. The new streamlined ODBC client improves the database performance from a client perspective by 3 to 5 percent. ODBC opens SQL Server access for non-Windows SAP application servers. SAP has postponed official support of using ODBC for non-Windows SAP application servers.

Data Compression Data compression in SQL Server 2008 includes: Storing numeric data more efficiently, such as int, bigint, float, and decimal data types Shrinking the size of tables that have primarily numeric data while optimizing the storage of default values Providing improved query performance in input/output (I/O) bottleneck scenarios Reducing hardware requirements and operation costs for storage and backup infrastructure Enabling compression at the table level

With SQL Server 2008 there are two types of data compression available: Row Compression A newly defined row format to optimize storing numeric values. ROW compression, as delivered in SQL Server 2008, is an extension of what was delivered for the specific decimal case in SQL Server 2005 SP2. Page Dictionary Compression Works on a per page level to store common prefixes and values only once on the page.

Figure 2 shows row compression with SQL Server 2008, which extends the variable length storage format of a decimal data type to other data types such int, bigint, and float. Other optimizations also reduce the space needed to store values like 0 or 0.0. Instead of consuming two bytes as in vardecimal, such a value will only consume four bits in SQL Server 2008.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 9 of 22

Figure 2. Row Compression with SQL Server 2008

The row compression feature provides a new storage format for a table that saves space without using a compression algorithm. A new row format has been introduced to help save more space, but it is not related to compression algorithms trying to compress a row on a page. Figure 3 shows page dictionary compression in SQL Server 2008, where a prefix list is stored in the page for common prefixes. Individual values are replaced by a token for the prefix, and a suffix for the value.
Figure 3. Page Dictionary Compression with Individual Values in SQL Server 2008 DateID 20070601 20070601 20070602 20070602 20070602 20070602 20070603 CarrierTracking 4911-403C-98 4911-403C-99 6431 6431-4D57-83 6431-4D57-84 6431-4D57-85 4E0A-4F89-AE OfferID 10 10 10 10 10 10 10 PriceDisc 0.00 0.00 0.00 0.00 0.00 100.00 0.00

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 10 of 22

Page dictionary compression has two optimizations, which apply to the data/index page of SQL Server. The first optimization helps find prefixes which are common in different rows. For example, the DateID column, which represents the way SAP stores dates in a varchar column, has a common prefix of 2007060 for all the values in the column DateID. SQL Server page compression stores this prefix value separately on the page and places a 12 byte token as a placeholder pointing to the shared prefix. Figure 4 shows page dictionary compression in SQL Server 2008, where a common value dictionary is stored in the page. Common values are replaced by tokens.
Figure 4. Page Dictionary Compression with Common Values in SQL Server 2008

In the second optimization the common values are replaced by a token, but only if space can be saved.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 11 of 22

Data Compression in SQL Server 2008 and 2008 R2 Compression Recommendations All customers are advised to implement SQL Server page compression on tables and secondary indexes on development, QAS, and production systems. Modern Intel and AMD-based servers are powerful enough to process the compression algorithms without any observed latency, and the reduction in I/O and improved cache hit ratios results in a performance benefit for end users. It is therefore recommended to enable page compression for all tables and secondary indexes on all SAP ABAP-based systems. New SAP installations have already been defaulted to use page compression for all tables and secondary indexes. SAP has changed the default compression type from row to page compression with the following results with SQL Server 2008 R2: Up to 80 percent reduction in database size on ECC Unicode Systems Approximately 70 percent reduction in database size on ECC Non-Unicode Systems Generally 80 percent reduction in database size on SAP components such as BW, CRM, and SCM Approximately 70 to 80 percent reduction in database size for customers moving from Oracle to SQL Server Virtually all customer deployments have shown significant improvement in performance ranging from 5 to 35 percent Negligible amount of measurable higher CPU consumption and in some cases a reduction in CPU consumption

For more information on page compression, see the Running Sap Applications on SQL Server blog at http://blogs.msdn.com/b/saponsqlserver/archive/2010/10/08/compressing-an-sap-database-using-reportmsscompress.aspx and the Data Compression: Strategy, Capacity Planning and Best Practices white paper at http://msdn.microsoft.com/en-us/library/dd894051.aspx. Star Join Query Processing SQL Server 2008 introduced a new way to build query plans around star or snowflake schemas. SQL Server joins dimensions with the fact table or the intermediate result set of the former join result, one dimension at a time. However, SQL Server rarely analyzes several dimension tables first, overlaying the matching rows and then accessing the fact table to perform a first join. This is demonstrated in the next sequence. A typical plan as observed today looks like joining the dimensions. This helps reduce the optimal number rows as with the fact table, and then takes the intermediate result set and joins that with the next dimension.
Figure 5. Star Join Query Processing

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 12 of 22

SAP 7.0, SQL Server 2000, and SQL Server 2005 build one dynamic bitmap filter, which is usually used for the first join.
Figure 6. Star Join Query Processing with One Dynamic Bitmap Filter

SQL Server 2008 allows you to build multiple dynamic bitmap filters.
Figure 7. Star Join Query Processing with Multiple Dynamic Bitmap Filters

These dynamic bitmap filters can be moved and overlaid before accessing the fact table. This means a combined filter of multiple dimension tables is applied in one step.
Figure 8. Star Join Query Processing with Moving Dynamic Bitmap Filters

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 13 of 22

Star join query processing can only be utilized with a maximum degree of parallelism greater than 1 (that is, star join requires at least a maximum degree of parallelism of 2). The standard recommended global maximum degree of parallelism must always be set to 1 for SAP systems. Microsoft and SAP have implemented a new parallelization setting that can be switched on for specific BW queries on cubes. These changes are implemented in SAP support packs or SAP note 1126568. For more information, see http://blogs.msdn.com/b/saponsqlserver/archive/2010/12/03/sap-bw-7-and-sql-server-2008-star-joinoptimization.aspx.

Transparent Features of SQL Server 2008

Backup Compression
Customer demand for backup compression is very high. So far, there have been only third-party solutions for SQL Server backup compression. Motivation for most of the customers asking for such a feature is the fact that backups should be kept on disk and/or that such backups should be transferred to centralized backup infrastructures. With SQL Server 2008, the capability of compressing any type of SQL Server backup comes out of the box. SQL Server 2008 offers the following advantages: Has WITH COMPRESSION clause to BACKUP backup database E61 to DISK = N'D:\E61full_compress.bak' with COMPRESSION Less storage required to keep backups online Backups run significantly faster, with less I/O Restore automatically detects compression and adjusts accordingly Integrated into Resource Governor to throttle CPU consumption Can be used by SQL Server log shipping functionality Tests so far reveal parity with LiteSpeed for SQL Server (from Quest Software) Usable by all third-party backup vendors through a VDI or VSS interface

Unlike LiteSpeed for SQL Server, SQL Server 2008 backup compression does not have any compression grades to configure. It is one grade only, which is possible with SQL Server. A comparison between different grades of LiteSpeed for SQL Server and SQL Server 2008 backup compression with a 5.5 terabyte SAP Unicode ERP shows the following: SQL Server 2008 backup compression provides a better compression rate with slightly more CPU consumption than LiteSpeed for SQL Server (which is configured for compression grade 1) LiteSpeed for SQL Server provides a slightly better compression rate with compression grade 11 and significantly more CPU resource consumption than SQL Server 2008 Backup Compression The transaction log compression rate is between 2.0 and 2.5. The compression rate on SAP ERP is between 4.0 and 5.0.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 14 of 22

Audit
Auditing, a new feature in SQL Server 2008, addresses the auditing of user activity, including compliance and auditing concerns for unapproved SAP users who are accessing the database directly. It helps to monitor the activities of system operators and SAP database administrators (DBAs). To achieve this, the new audit functionality was created with a first-class server object. This means the auditing is not done using SQL Server Profiler, but by defining audit sessions and audit rules through simple T-SQL statements. No additional tools are necessary. This functionality is completely integrated into the SQL Server relational engine.
Figure 9. Capabilities of Different Tools Moved into one SQL Server Engine Functionality to Audit Activities

To make auditing of non-SAP motivated database accesses successful in a SAP landscape, you should adhere to the following guidelines: Never keep the log on the database server Do not give SAPServiceSID and SIDAdm access to the share on the server where the log resides Do not give any DBAs access to the share on where the log resides Delete the BUILTIN\Administrator Login in SQL Server SQL Server should run in the context of an account, which is not allowed to login interactively

While it is not always possible to prevent a system administrator from disabling the audit session or changing its definition, you can build an audit session that tracks: Who logged into SQL Server and got mapped into the system administrator role What changes were done to the audit session with that logged in session

Although the very same person does not have access to the audit log, it should not be possible for that person to manipulate the results of the audit log showing manipulations.
Figure 10. Simple Audit Session and its Specifications Defined

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 15 of 22

Transparent Data Encryption

SQL Server introduced the encryption of columns within a database with SQL Server 2005. However, this encryption feature was not transparent to applications. In the case of some applications, significant changes were required to leverage this feature. It also had the disadvantage of not being able to leverage index searches on such columns. In SQL Server 2008, a new functionality has been introduced called Transparent Data Encryption. With this the scope of encryption is now extended to the whole database. The encryption itself will be transparent to an application since accessing or not accessing an encrypted database is now at the SQL Server instance. That means that once a SQL Server instance opens an encrypted database, every application or user can access the data in this database. This also means that encryption is not a mechanism to control user access or to prevent some users from accessing the content. This is an important difference from the column-level encryption introduced with SQL Server 2005 which allows such access controls. The encryption/decryption of transparent data encryption is done in the I/O path. This means that the data is in a non-encrypted state in the SQL Server buffer pool and can fully be leveraged for indexing and index searches. Transparent data encryption includes the following capabilities: Four different encryption algorithms, including AES_128, AES_196, AES_256, and TRIPLE_DES Encryption of all new pages written to disk and triggering of background threads to encrypt the rest of the database Improved throughput and reduced overhead with AES algorithms Not recommended to use TRIPLE_DES since it is the weakest encryption algorithm and is not as efficient in resource consumption as the AES algorithms Overall impact hardly measurable Backing up an encrypted database will automatically provide an encrypted backup Negligible backup compression achievable on an encrypted database using SQL Server backup compression or LiteSpeed for SQL Server Transparent data encryption and SSL can be used in conjunction (For more information, see SAP note 1570930) Transparent data encryption, backup compression, SSL, database compression and other technologies can be used in conjunction (For example, SQL Server fully supports a Transparent Data Encryption database that is fully PAGE compressed to use backup compression and SSL.) SAP specific white paper will be released before SQL Server 2008 RTM (For more details, see OSS note 1380493.)

Encrypt Connections using Secure Sockets Layer

Customers in the defense, security, public safety, government, and financial industries sometimes require additional protection from man-in-the-middle and other related security threats. SQL Server, like other commercially available databases, supports encrypting the communication channel between SQL Server and the client application using the Secure Sockets Layer (SSL). The encryption mechanism is handled between the SQL Native Access Client (SNAC) and SQL Server and is therefore transparent to the application layer. For more information on how to setup SSL network encryption between SQL Server and SAP, see http://blogs.msdn.com/b/saponsqlserver/archive/2011/03/09/sql-server-network-encryption-with-sap.aspx and OSS note 1570930.
SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com
Page 16 of 22

Lock Escalation Granularity

The locking strategy of SQL Server is called dynamic locking. SQL Server in its relational engine tries to be smart in terms of using a specific lock granularity. SQL Server is built to avoid allowing lock structures to take over a significant amount space in the buffer pool. Therefore, SQL Server versions released so far have a few events where the lock granularity could be escalated to a higher granularity. The events show the following: 40 percent buffer pool used by lock structures. This usually means a few million locks in the SAP space. More than 5,000 row locks held by a DML statement on one table. In this case, there is an attempt to escalate to table lock granularity. If there is another transaction with other locks on the table, one will continue by acquiring row-level locks. This type of escalation attempt can be suppressed by trace flag 1211 or 1224. Storage Engine starts a statement with page-level locks, and it cannot be disabled with sp_indexoption. Storage Engine decides to start a statement with table-level locks, and it cannot be disabled.

An issue with lock escalation granularity is the lack of a central switch to indicate that row level locks for a specific table should always be used, with a command such as: ALTER TABLE VBDATA set (LOCK_ESCALATION = DISABLE).

Minimal Logging Enhancements

SQL Server 2008 introduced several changes around the way logging is done. The most significant changes have affected Unicode or platform migrations of SAP applications. Too much transaction log space is consumed with Unicode migrations involving databases with a large volume of terabytes. The series of events leading to this higher consumption of transaction log space includes: Loading tables in a SAP installation or Unicode/platform migration that includes: Creating empty table Creating a clustered index Loading data in batches of 10,000 rows Creating non-clustered indexes

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 17 of 22

These events lead to the following: Only the first batch is minimally logged All subsequent batches get fully logged, which means that millions of rows are entered into the transaction log as well A large volume is required in the SQL Server Transaction Log and it takes up too much space, especially in Unicode/platform migrations

SQL Server 2008 expands the minimal logging behavior to all batches loaded. This has cut down the required space for the SQL Server transaction log by a factor of up to 50 in some observed instances.

Query Plan Freezing

In SQL Server 2005, Microsoft introduced a feature called plan guides. This feature should have helped to get around one of the paradigms current databases have. A database management system needs to cache access plans (query plans) since generating these in real time cannot be done with 5,000 or 10,000 statements being executed per second. In contrast, side skewed data or extremely uneven data distribution in some columns of tables does not guarantee that an access plan created for one set of values submitted to a query fits perfectly for a different set of values. The idea was to generate an access plan from the outside without having to change the application by introducing query hints. However, plan guides have not been used frequently in the SAP space since customers found difficulty using the feature. SQL Server 2008 improves on plan guides and provides a way to create plan guides or better freeze existing query plans based on plan_handles, which can find a DMV called sys.dm_exec_query_stats. For more information and queries, see http://blogs.msdn.com/saponsqlserver/archive/2007/06/05/sap-dbacockpit-and-some-related-sqlscripts-part-1.aspx. The typical use case in the SAP space for this feature is: Not to allow plan changes for critical plans Troubleshooting queries (cardinality tuner) Plan fixing for queries accessing tables/columns with extremely skewed data

Query Plan Freezing works as follows: 1. 2. 3. 4. Checks sys.dm_exec_query_stats to find the plan to be frozen Finds queries to freeze at http://blogs.msdn.com/saponsqlserver/archive/2007/06/05/sap-dbacockpit-andsome-related-sql-scripts-part-1.aspx Takes a plan_handle and checks whether the query has the plan to be frozen Takes sql_handle and statement offset and executes sp_create_plan_guide_from_cache @name = Nmyfirst_plan_guide', @plan_handle = <sql_handle>, @statement_start_offset = <statement offset>

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 18 of 22

SQL Server Database Mirroring Enhancements Compression

SQL Server database mirroring is being used more frequently in the SAP space, especially by larger companies. In SQL Server 2008, Microsoft made small but meaningful changes to database mirroring. The first enhancement includes compressing transferred data after it has been encrypted, which significantly reduces the network bandwidth requirement.
Figure 11. SQL Server 2008 Database Mirroring Enhancements Using Compressed Log Records

SQL Server Database Mirroring Enhancements AutoRepair

AutoRepair is another new feature added to SQL Server 2008. This functionality catches cases where SQL Server detects a consistency issue on one or multiple pages. In the case of synchronous mirroring, database mirroring lets one database check another database to see if the page in question is healthy. If this is true, database mirroring is exchanging the inconsistent page of one system with the healthy page from the other system. AutoRepair works by allowing the principal to detect an inconsistent page in its database to get the healthy page from the mirror, or vice versa. This means that the database mirroring configuration needs to be synchronous. AutoRepair does not work with asynchronous database mirroring.
Figure 12. AutoRepair with Synchronous Mirroring

For more information on database mirroring, see OSS note 965908, SQL Server Database Mirroring and SAP Applications.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 19 of 22

Changed Autogrowth Behavior

Autogrowth behavior does not work well with the way SAP applications are ported to SQL Server. Figure 13 demonstrates the problem with SAP applications ported to SQL Server:
Figure 13. Using Trace Flag 1117 to Adjust Autogrowth Behavior All the data files have the same size and the files are getting filled in even portions.

The problem occurs when all the files are running full. In this case, the current algorithms will grow by one file at a time. This kind of autogrowth behavior destroys the even fill that is ideal for different data files.

SQL Server 2008 introduces startup trace flag 1117. This trace flag changes the autogrowth behavior to one where all data files are grown in the same size (assuming the autogrowth rate is set the same for each of the files), instead of growing one data file only.

For more information on trace flag 1117, see OSS note 1238993, Proportional File Auto-Growth with SQL Server 2008.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 20 of 22

Appendix
SAP and SQL Resources
http://www.sdn.sap.com/irj/sdn/mss http://www.microsoft.com/isv/sap/technology/platform/sql.aspx http://blogs.msdn.com/saponsqlserver/ http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/50c5de8f-282a-2e10-cb8aa5859a0994a4?QuickLink=index&overridelayout=true

SAP Notes
This section includes SAP notes on a wide variety of topics relevant to SQL Server technologies for SAP solutions. 1006887 - Online index creation with SQL Server 2005, 2008 and 2008 R2 1067103 - Transaction context lost 1076022 - Release planning for Microsoft SQL Server 2008 (R2) 1126568 - Enable star join optimization with SQL Server 2008 1134345 - Using locked pages for SQL Server 1141630 - Preparations for SQL 2008 BW customers 1144459 - SQL4SAP and SAP Installation Media for SQL Server 2008 1152240 - Setting Up Microsoft SQL Server 2008 1158936 - DBACOCKPIT Backup compression option in SQL 2008 1174635 - TempDB sizing in SAP BW systems on Microsoft SQL Server 1177356 - MS SQL Server End of Support for SAP Releases 1237682 - Configuration Parameters for SQL Server 2008 1238993 - Proportional File Auto-Growth with SQL Server 2008 1241751 - SQL Server minimal logging extensions 1252970 - Triggers on SAP tables 1280476 - SQL 2008 Cluster The current SKU is invalid 133381 - Database-Hints in Open SQL for MS SQL Server 1380493 - SQL Server Transparent Data Encryption (TDE) 142731 - DBCC checks of SQL server 1443424 - Migration path to Win2008MSSQL2008 for 4.6C and 6.206.40 1459005 - Enabling index compression for SQL Server 1476928 - System copy of SAP systems on Windows 2008 (R2) SQL Server 1488135 - Database compression for SQL Server 1494789 - Enabling 1000+ partitions support on SQL Server 1497623 - SQL Server Database Files Should Not be Sparse 1505884 - R3load compression for table index 1570930 SQL Server network encryption with SAP 151603 - Copying an SQL Server database

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 21 of 22

159316 - Reorganizing tables on SQL Server SQL 20002005 208632 - TCPIP network protocol for MSSQL 28667 - Microsoft SQL Server Specific Profile Parameters 32129 - Deadlock analysis for the SQL server 327494 - Configuration Parameters for SQL Server 2000 363018 - File management for SQL Server 392892 - DBIF_RSQL_SQL_ERROR (SQL error 0 or 11) 398136 - Support Policy for Microsoft SQL Server 493290 - Configuring SQL server Log shipping 521750 - FAQ SQL Server 2000 IO performance 551915 - R3 won't start after database restore or database copy 592514 - MSSQL internal database users, permissions and security 600027 - Installing the Corrected MS SQL Server Collation 62988 - Service packs for Microsoft SQL Server 652634 - FOR ALL ENTRIES performance with Microsoft SQL Server 683447 - SAP Tools for MS SQL Server 799058 - Setting Up Microsoft SQL Server 2005 849062 - Optimizer statistics for InfoCubes in BW 869407 - Partitioning on MS SQL Server 2005 879941 - Configuration Parameters for SQL Server 2005 905634 - Release planning for Microsoft SQL Server 2005 947702 - DB13 - DBCC Checkdb job fails in SQL 2005 SP1 960769 - Windows Migration from 32-bit to 64-bit (x86_64) 965145 - Installation of 46C SAP Systems on MS SQL Server 2005 965908 - SQL Server Database Mirroring and SAP Applications 985137 - Service Pack Installation for SQL Server 2005 98678 SQL Server Connection Issues 987961 - FAQ SQL Server 2005 IO performance 989595 - FAQ Cache Hit Ratio in SQL Server 991014 - Vardecimal Compression for SQL Server 2005

Microsoft Blogs
http://blogs.msdn.com/saponsqlserver/ http://blogs.msdn.com/saptech/ http://blogs.technet.com/virtualization/ http://blogs.msdn.com/b/saponsqlserver/archive/2011/03/18/increased-partition-support-of-sql-server-insap-bw.aspx

The SAP SD Standard Application Benchmark Results for Two-Tier Internet Configuration: SAP enhancement package 4 for SAP ERP 6.0; SQL Server 2008 and Windows Server 2008 R2 Enterprise Edition; 4 processors, 40 cores, 80 threads, Intel Xeon Processor E7-4870, 2.4 Ghz, 64 KB L1 cache and 256 KB L2 cache per core, 30 MB L3 cache per processor. For more details, see certification number 2011016 at http://www.sap.com/solutions/benchmark/sd2tier.epx.

SAP DEVELOPER NETWORK | sdn.sap.com 2011 SAP AG

BUSINESS PROCESS EXPERT COMMUNITY | bpx.sap.com

Page 22 of 22