how sap's vision comes to life (pdf)

16
SAP Sybase Adaptive Server Enterprise What’s New in ASE 15.7 SP100 © 2013

Upload: vothuan

Post on 11-Dec-2016

246 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: How SAP's Vision Comes to Life (PDF)

SAP Sybase Adaptive Server Enterprise What’s New in ASE 15.7 SP100 © 2013

Page 2: How SAP's Vision Comes to Life (PDF)

TABLE OF CONTENTS Introduction ................................................................................................................................................. 3

Introducing ASE 15.7 SP100 ...................................................................................................................... 4

VLDB Performance ..................................................................................................................................... 4

Query Performance ..................................................................................................................................... 6

Operational scalability ................................................................................................................................ 6

Storage Optimization and Recovery ......................................................................................................... 8

Integrated DR solution for Planned Downtime ........................................................................................ 9

Monitoring and Diagnostics ..................................................................................................................... 11

SAP Control Center (SCC) ASEMAP Enhancements ............................................................................ 13

Other Useful Functionality ....................................................................................................................... 14

Summary .................................................................................................................................................... 15

Page 3: How SAP's Vision Comes to Life (PDF)

Introduction In today’s fast-paced and fluctuating economy, organizations are striving to provide customers with better

products and services, optimize their operations and fit in the need to act fast within an ever-shortening

maintenance and decision window. Success is measured in the form of better performance combined with

providing the right information at the right time with the right degree of accuracy.

SAP offers a versatile high performance enterprise-class relational database management system that is

well suited for mission-critical, data-intensive environments. It ensures highest operational efficiency and

throughput on a broad range of platforms offering the lowest cost of ownership (TCO) of enterprise-class systems.

Adaptive Server Enterprise from SAP has made tremendous strides forward in improving operational efficiency,

increasing security, exploiting parallelism to provide significant performance gains including enhancements to

provide significant storage optimization and allowing ease of monitoring and development helping not only

database administrators in efficiently managing their database servers but also the CIOs in delivering business

insights and value cost-effectively.

. ASE 15.7 ESD2 had introduced functionality that would help address customer needs including VLDB

performance, security and monitoring:

VLDB Performance

o Online Utilities

o Split / Merge Partitions

o Fast Logged BCP

o Expanded DB Size

o Pre-computed result sets

o Hash based update statistics

o User Defined Opt Goals

o Query Plan Sharing

o Enhanced Dump and Load

o Concurrent Dump DB and Dump Tran

o Async DB initialization

o Deferred Table creation

o Alter table drop column without data copy

o In-row large object compression

Page 4: How SAP's Vision Comes to Life (PDF)

Security

o Granular Permission Control

o Predicated Privileges

o Support for Encryption of Text/Image/Unitext

Monitoring and diagnostics

o Show Cached Plan in XML

o Compressing Shared Memory Dumps

SAP Adaptive Server Enterprise 15.7 SP100 is focused on operational scalability along with providing

critical functionality that will enhance the performance and improve monitoring.

Introducing ASE 15.7 SP100 SAP Sybase Adaptive Server Enterprise 15.7. SP100 is the latest pillar to the ASE 15.x series which was

designed to address the challenges of terabyte-scale operational environments.

In this article, we will look at some of the exciting new functionalities and capabilities that will benefit

customers once they upgrade to ASE 15.7 SP100. Some of the key enhancements that will help ASE 15.7 SP100 in

meeting the following goals for this release are described in the sections below:

Improved VLDB Performance

Improved Operational Scalability

Improved monitoring

VLDB Performance Create Index Performance Enhancement

It is now possible to issue a create index statement to be executed in a parallel manner resulting in

shorter index creation time. This enhanced parallel ‘create index’ is disabled by default and is part of the ‘enable

functionality group’.

In order to make use of the ‘parallel’ feature of ‘create index’, the following will have to be executed:

1) sp_configure “enable functionality group”, 1

2) sp_dboption <database_name>, “select into”, true

And then the following ASE configuration parameters have to be set according to hardware environment:

Page 5: How SAP's Vision Comes to Life (PDF)

1) number of worker processes – to set to the maximum number of concurrently executing parallel

threads used by all users

2) max parallel degree – to be set to the maximum parallel degree used for an individual user but not

higher than ‘number of worker processes’

3) max online engines – SAP recommends configuring sufficient ‘number of engines’ when configuring

parallel threads depending on the hardware availability and other workloads. ‘number of worker

processes’ is typically set higher than ‘number of engines’

The basic requirements for ASE to select an enhanced parallel query plan are:

1) The table where the index is created uses a data-only-locked format and is not partitioned or empty.

2) The index is not a non-clustered index

3) The leading column of the index has at least two distinct values.

Create index with Hash Stats

In most database environments, indexes are created to efficiently manage data access leading to

performance improvements during query execution. In general, executing ‘create index’ on large tables is bound

to have a performance impact especially gathering index statistics. Adaptive Server 15.7 SP100 extends the

support of hash-based statistics gathering to the create index command. ‘Create index’ command will now have a

‘with statistics’ extension to support hash-based stats.

The create index command with hash-based statistics gathering options allow minor index attributes to

be gathered while the index is being created serially or in parallel. The biggest advantage to this functionality is

that users will not have to execute ‘update statistics’ after the ‘create index’ command with regards to minor index

attributes and significant performance gain can be expected.

ODBC : Send Client Attributes with next exec

In an application environment, LOB data is handled by the application by executing SQLBulkOperations

and by using ODBC batching protocols like SQLExecute/SQLExecDirect. Since data-at-exect is not supported, LOB

data must be fully materialized before executing these protocols.

A new API called ‘data-at-exec’ is being provided which will help the application to avoid materializing the

LOB parameters prior to the call to SQLExecute. Instead LOB data will be sent in chunks and the ODBC driver will

convert the data as it is received and hold on to it until all data for the parameter is received at which time it will

be sent to the server.

Page 6: How SAP's Vision Comes to Life (PDF)

Query Performance Enhance star joins

Adaptive Server 15.7 SP100 introduces a new query plan optimization feature called the ‘bloom filter’,

which is used to improve join performance. A bloom filter provides early filtering of non-joinable rows before they

would reach the join operator. Bloom filter is only considered when one child of the join operator has to

materialize its derived table before the other child is opened. ‘Bloom filter’ provides a low-cost approach to

eliminate rows before they participate in a full join algorithm.

In Adaptive Server, ‘bloom filter’ is implemented for hash join, sort merge join and reformatting based

nested loop join. ‘The bloom filter’ is only enabled when optimization criteria statement join_bloom_filter is ON.

Optimizer decides to use ‘bloom filter’ in the final plan if the usage of the bloom filter is estimated to reduce the

total cost of the join.

Operational scalability Incremental Reorg

In the database world, maintenance windows are very critical to ensure operational scalability. Typically

“reorg” utility helps in defragmenting the data and improves access performance. Adaptive Server 15.7 SP100

adds a parameter defrag to the reorg command, allowing the user to schedule and resume reorganization while

also allowing concurrent reads or writes on the data being reorganized.

This feature is focused on improving the capabilities of “reorg online” where reorg could be done without

blocking regular operations. “Reorg defrag” now splits the reorganization event into multiple transactions one

right after the other compared with the traditional reorg which reorganizes the entire table in a single transaction.

This allows the user to specify a time interval to reorganize the data and resume the process at a later time until

completion. The advantage to this is the ability to keep the maintenance window small and improve productivity

thereby increasing data availability and reducing data disruption to business operations. Here is an example of

using the “defrag” command where we will be performing incremental reorg for 3 minutes

reorg DEFRAG demotable with TIME=00:03 go

Page 7: How SAP's Vision Comes to Life (PDF)

Ability to Create Index Online

It is becoming more and more important for businesses to have high data availability. Traditionally,

creating an index on a large data has been known to cause a significant performance problem in terms of data

availability. Adaptive Server 15.7 SP100 includes the create index ... online parameter providing the ability to

create index online where index creation will not block the data availability. All parameters associated with the

existing ‘create index’ command will work the same with the “online” option except the ‘sorted_data’ parameter

which creates index on the existing data layer.

Support for Parallel Index Update

As the data size increases, there is an impact on the performance of loading of data into the tables. This

feature has been implemented with the goal to improve the data load performance when the target table has

multiple indices defined on it and the mass data load is accomplished by an insert statement.

Insert-Select Performance

Performance has played a very important role for ASE in comparison with competition. The performance

improvements that has been addressed in ASE 15.7 SP100 with regards to massive data insert during the loading

process has solidified the fact that ASE is continuously improving in the area of performance. Insert…Select

performance is improved by applying bulk insert mode. ASE 15.7 SP100 introduces a new optimizer criteria

“INS_BY_BULK” to the bulk insert mode and this can be applied at a system, session and individual query level.

One of the advantages of using this approach is that the user applications do not have to be changed in order to

use this feature. An additional benefit that comes for free as part of this enhancement is the improvement to

performance of statements like “Select Into” and worktable population. Thus, this optimization helps to improve

performance during massive loads into a table.

Page 8: How SAP's Vision Comes to Life (PDF)

Storage Optimization and Recovery Shrink database / drop devices

ASE 15.7 SP100 introduces the concept of shrinking the database which is a sought-after functionality in

the customer base. This feature allows freeing of unused space for reuse and/or deletion. ASE removes the

associated physical storage from the database without losing the data located on that storage. ASE moves the

data on the portion of the database to be shrunk to a new location before attempting to remove the space from

the database. Once the portions to be removed are empty, their physical storage will be replaced by references to

a null device thus making the freed space available for reuse. ASE uses the “alter database” command to shrink

the database. Here is an example of shrinking a database by removing a database device from the database:

alter database demoDB OFF dbdev2

The OFF clause specifies the device name(s) from which space is to be released. It cannot appear in a command

together with either the ON or LOG ON clauses.

Incremental dump database

One of the responsibilities of a DBA is to ensure that the database is backed up and the latest dumps are

available in case there is a need to recover from a crash or a similar situation. ASE 15.7 SP100 has included some

enhancements to the dump and load commands that will keep the DBAs with a secure feeling about their database

environment. The dump database command now allows a cumulative backup to be performed where one can

dump only those database changes since the last full database dump. Another enhancement to the backup

functionality is the ability to dump one database and load it into another database using the dump history file. It

should be noted that the Tivoli Storage Manager is now fully integrated with the dump history file. A new utility

called sybdumptran helps in generating a final transaction log dump from log devices from outside of ASE in case

of a catastrophic failure. These enhancements bring several advantages like reducing the amount of time needed

for a full backup, speeding up recovery time, improving backup performance etc.

Page 9: How SAP's Vision Comes to Life (PDF)

Enhanced Dump and Load

Adaptive Server 15.7 SP100 has now included a new standalone utility called ‘restore’, which gives the

ability to restore a database from the database and transaction log dumps. In order to support this utility, the

load database and the load transaction commands have been enhanced with new options. With the new changes,

it is now possible to apply source database attributes to the target database. It is essential that both the target

and the source database have the same attributes to ensure that the restored database is exactly as it was.

Restore Command Line Tool

The requirement is to have a stand-alone utility to restore an ASE database to the time of failure from the most

current full database backup dump files. In order to use the utility, the ASE server must be up, the master

database is available, source database is connected to the ASE server in a recovered state or unrecovered state

and the log segment of the source database should be available for dumping and then loading back the last

transaction logs that have not been dumped. Ideally the utility should enable user to restore from the log segment

of the database for the transactions not backed up to the dump files in case of data disk crashes. The utility should

provide a simple command line interface as well as a silenced “parameterized” interface.

Integrated DR solution for Planned Downtime Support of writetext command to replicate LOB changes

If a database had tables with LOB columns, the process of marking the database for replication included a

step where the database server maintained a storage area that will be updated with the information about the off-

row columns or LOB columns. When the shrinkdb operation moves information around to accommodate the

reduction in physical storage, it needs a way to locate the data row pointing to it prior to the shrink operation.

Earlier, only when replication is enabled, the row-id pointing to the home data row gets updated with the new

page number. With this release, support has been added to maintain the information of the LOB columns so that

the operations like shrinkdb will not affect the link to the home data row.

Page 10: How SAP's Vision Comes to Life (PDF)

Provide soft quiesce functionality

In a typical HA/DR environment, there will be two or more servers one of which will be the primary where all the

transaction processing occurs while the rest will serve as warm standby to the primary. An HA/DR configuration

offers a seamless planned application failover from primary to standby.

A new functionality called ‘Soft Quiesce’ is introduced in this release which will ensure that the transaction

activities are terminated in a proper sequence on the primary, followed by the draining of transaction logs of the

HADR databases by the replication subsystem in case of a disaster. In order to support a seamless planned

failover, it is essential to stop transaction processing activities on HADR databases on the primary in an orderly way

so that no further changes can occur in the transaction logs. This is the quiesce state of the log. At this point, the

replication agent thread using the LOG DRAIN mechanism drains the transaction log. Once the draining is

complete, the inactive primary is ready for failover.

MRP support for multiple scanners

The Adaptive Server monitoring tables have been enhanced to support multiple scanner threads for RepAgent.

Enabling multiple scanner threads for RepAgent provides access to available replication paths and improved

replication performance. It also helps the Adaptive Server monitoring tables to provide information about

RepAgent scanner task performance. The monitoring table ‘monRepSenders’ now has three columns that have

beed added to provide compatibility with the single task model. The three columns that have been added are:

• NumberOfTruncPointRequested

• NumberOfTruncPointMoved

• AvgTruncPointInterval

• monRepScanners

The NumberOfTruncPointRequested and NumberOfTruncPointMoved columns are populated when the single task

scanning of a log model is used. When the Multiple Replication Paths (MRP) model is used, these columns values

are 0.

Page 11: How SAP's Vision Comes to Life (PDF)

Monitoring and Diagnostics Server-wide DML metrics

With the increased requirement for better performance, there is a need for being able to provide metrics on a

number of operations reflecting two different areas:

1) Application activities such as committed and rolled back transactions, select, update, insert, delete or

merge queries.

2) Utilization of server resources such as table scans, index scans, temporary and work tables, user log

cache activities, I/O activities such as physical reads and writes and logical reads.

These metrics help the end-user get a detailed understanding of the activity that has occurred as well as those that

are occurring currently on ASE.

Elapsed time monitoring

Although it is possible to obtain detailed information for how much time ASE is spending waiting for events such as

locks, I/O operations or a wide range of internal events, there was no visibility into the total time taken by the

server in performing the actual work to get this information. The following execution categories have been

enhanced to track the number of times the operation has been performed and the total amount of time spent

performing operations :

- Query execution

- Sorting

- Query compilation

Support built-in grouping of SQL Statements

Adaptive Server 15.7 SP100 includes the show_condensed_text function, which gives the ability to

aggregate monitoring data for multiple statements in the statement cache which, although they are syntactically

distinct (having individual comments or differences in their text format), represent semantically identical queries.

This capability allows the function to aggregate the values from the MDA table monCachedStatement into a single

row in the query result. The information thus obtained is helpful when making a decision on which statements to

retain in the cache.

The MDA table monCachedStatement stores detailed monitoring information about the statement cache,

including, but not limited to:

• Information about resources used during the previous executions of a statement,

Page 12: How SAP's Vision Comes to Life (PDF)

• How frequently a statement is executed

• The settings in effect for a particular plan,

• The number of concurrent uses of a statement

Query Plan and Execution metrics output in HTML format

In Adaptive Server versions 15.7 SP100 it is possible to have a graphical query plan generated in HTML

format for viewing in a Web browser. Earlier, one method of performing analysis of complex queries was to

evaluate text based output provided by the set showplan and set statistics {time | IO} commands. The HTML

representations are displayed on a single page and the description of the query plans are more intuitive and

clearer.

Access in-use plans through show_cached_planIin_xml()

ASE 15.7 ESD#2 introduced a built-in function to provide access to the query plans for cached statements

in XML format. However, there was a restriction that prevented it from returning the plans for query plans that

were in use at the time that the function was executed. With Adaptive Server versions 15.7 SP100 the restriction

has been removed and the plans for query plans will now be returned.

Enhance monSpinlockActivity MDA Table

The MDA table monSpinlockActivity captures information on the spinlocks defined in the running server.

The enhancement that is being done to the MDA table monSpinlockActivity in ASE 15.7 SP100 introduces a new

column that will contain a slot ID associated with each spinlock. The new column is called SpinlockSlotID in order

to indicate the value of the column that is derived from the Object Pool slot in which the spinlock memory is

allocated and is not a persistent property of the spinlock itself. This provides a detailed analysis of the usage of

individual spinlocks in the server.

Page 13: How SAP's Vision Comes to Life (PDF)

SAP Control Center (SCC) ASEMAP Enhancements SCC has included full support for all the important ASE 15.7 SP100 features thus making it easier

for the end users. Here are a few features for which SCC extends support in ASE 15.7 SP100:

- Cumulative Dump and Load

o SCC now supports the functionality of taking incremental database dumps as well as loading

these incremental dumps into a database.

- Enhancements to Dump and Load

o SCC is now being fortified with more enhancements to the dump and load commands.

o It is possible to generate the DDLs required to create a target database based on the

information of the source database. This is obtained from the dump history file that gets

recorded when ‘dump database’ and ‘dump tran’ commands are executed.

o As an extension, SCC can also generate the statements for ‘loading a database’ which is

done by reading the dump history file and generate the load sequence.

o Yet another enhancement is the ability to perform ‘load database’ from a point in time. SCC

also allows users to use multiple stripes when dump and load a database.

- Shrink Database

o SCC users can now determine if it is indeed necessary to execute the shrink database

command based on the space utilization information for the database.

- Incremental Reorg

o SCC now provides the ability to execute reorg operations on a table, index and partitions.

Using a wizard, it is possible to run reorg from the context menu of a database. These

wizards covers all the reorg options available in ASE to perform defragmentation.

- Create Index Online

o SAP control center allows a user to mark an index as online, in the index creation wizard,

indicating that data is not to be blocked while the index is being created.

- Support for Split/Merge/Move Partitions

o Split/Merge/Move partitions is a feature introduced in ASE 15.7 ESD#2. It is now possible to

execute split/merge/move partitions through SCC starting ASE 15.7 SP100. SCC supports the

ability to split a single data partition into multiple partitions, merge two data partitions into

one partition and move one partition from one segment to another.

Page 14: How SAP's Vision Comes to Life (PDF)

Other Useful Functionality ODATA (Open Data Protocol) Support

OData (Open Data Protocol) enables data services over RESTful HTTP. It provides the ability to

perform operations through URIs (Universal Resource Identifiers) to access and modify information. The Adaptive

Server OData Server consists of : the OData Producer and an HTTP server.

a) OData Producer – a Java Servlet that uses the JDBC API to connect to an Adaptive Server. The OData Producer

processes OData requests and responses and interfaces with the database.

b) An HTTP server - that handles OData requests from web clients – the OData server uses the Jetty WebServer as

its HTTP server. This embedded HTTP server also acts as a Java Servlet container, which is required to host the

OData Producer.

Instead of using the embedded HTTP server, it is possible to use a custom HTTP server to handle OData requests,

as long as the customer’s solution can also host Java servlets. For example, it is possible to set up an IIS or Apache

server to forward requests to a Tomcat server.

OData client requests are sent to an HTTP server through URIs and are processed by the OData Producer, which

then interfaces with the database server to issue database requests and retrieve content for the OData responses.

The OData schema for each client is based on the client's database connection permissions. Clients are not able to

view or modify database objects that they do not have permission to view.

Page 15: How SAP's Vision Comes to Life (PDF)

Summary Sybase ASE 15.7 SP100 is a recent addition to SAP ASE 15.7 with significant improvements to VLDB performance,

Operational Scalability, reliability using HA/DR solution and monitoring and diagnostics.

In the database world, increased data volume translates into various issues like operational scalability, decreased

performance and crunch in storage. SAP Sybase ASE 15.7 SP100 has introduced several key features which help

solve the issues that surround the increased data volume in an organization. From redistributing data for

maintaining partitioned data effectively to improving data availability during maintenance windows in a production

environment, ASE 15.7 SP100 has successfully enhanced the operational scalability experience for customers. In

the query processing area, ASE 15.7 SP100 introduces creation of index online, support for parallel index updates

and tremendous improvement in the area of insert-select performance supporting the load of massive data into

tables. Shrinkdb which is a storage saving feature and HA/DR for planned downtime in terms of reliability would

be well received by both new and current customers alike. Great enhancements made in the area of monitoring

and diagnostics would be very helpful for customers to monitor and communicate diagnostics information with

SAP.

SAP ASE 15.7 SP100 is available to customers as of May 2013.

Page 16: How SAP's Vision Comes to Life (PDF)

www.sap.com

© 2013 SAP AG. All rights reserved.

SAP, R/3, SAP NetWeaver, Duet, PartnerEdge, ByDesign, SAP

BusinessObjects Explorer, StreamWork, SAP HANA, and other SAP

products and services mentioned herein as well as their respective

logos are trademarks or registered trademarks of SAP AG in Germany

and other countries.

Business Objects and the Business Objects logo, BusinessObjects,

Crystal Reports, Crystal Decisions, Web Intelligence, Xcelsius, and

other Business Objects products and services mentioned herein as

well as their respective logos are trademarks or registered trademarks

of Business Objects Software Ltd. Business Objects is an SAP

company.

Sybase and Adaptive Server, iAnywhere, Sybase 365, SQL

Anywhere, and other Sybase products and services mentioned herein

as well as their respective logos are trademarks or registered

trademarks of Sybase Inc. Sybase is an SAP company.

Crossgate, m@gic EDDY, B2B 360°, and B2B 360° Services are

registered trademarks of Crossgate AG in Germany and other

countries. Crossgate is an SAP company.

All other product and service names mentioned are the trademarks of

their respective companies. Data contained in this document serves

informational purposes only. National product specifications may vary.

These materials are subject to change without notice. These materials

are provided by SAP AG and its affiliated companies ("SAP Group")

for informational purposes only, without representation or warranty of

any kind, and SAP Group shall not be liable for errors or omissions

with respect to the materials. The only warranties for SAP Group

products and services are those that are set forth in the express

warranty statements accompanying such products and services, if

any. Nothing herein should be construed as constituting an additional

warranty. .