Friday, June 15, 2018

Db2 for z/OS Performance Traces Part 3 - Trace Destinations and IFCIDs

In parts 1 and 2 of this series we examined each of the six types of Db2 traces that you can start to get Db2 to record information about its activities. Be sure to check out those earlier blog posts to learn about what is available. 

Today we are going to look at a couple of additional details with regard to Db2 traces, starting with where the trace records can be written.

Trace Destinations

When a trace is started, Db2 formats records containing the requested information. After the information is prepared, it must be externalized somewhere. Db2 traces can be written to six destinations:

  • GTF, or Generalized Trace Facility, is a component of z/OS and is used for storing large volumes of trace data.
  • SMF , or System Management Facility, a source of data collection used by MVS to accumulate information and measurements. This destination is the most common for DB2 traces.
  • SRV is a routine used primarily by IBM support personnel for servicing Db2.
  • OPn, where n is a value from 1 to 8, is an output buffer area used by the Db2 Instrumentation Facility Interface (IFI).
  • OPX is a generic output buffer. When used as a destination, OPX signals Db2 to assign the next available OPn buffer (OP1 to OP8).
The Instrumentation Facility Interface, which is a DB2 trace interface, enables Db2 programs to read, write, and create Db2 trace records and issue Db2 commands. Many online Db2 performance monitors are based on the IFI.

But which trace destination should you use? Well, it depends! Consult Table 1 below for a synopsis of the available and recommended destinations for each Db2 trace type. A "Y" indicates that the specified trace destination is valid for the given type of trace; Default indicates that it is the default destination for that trace type.

Note: you can click on the Table to expand its size for easier viewing.

If you use SMF as your Db2 trace destination, consider using the SMFCOMP system parameter (DSNZPARM), which was introduced in Db2 10. This parameter causes the compression of Db2 trace records written to SMF. When SMFCOMP is set ON, the z/OS compression service CSRCESRV is used to compress any Db2 trace records written to SMF.

Use this parameter if your shop is concerned about the volume of SMF records written by Db2.

Using IFCIDs

Sometimes a trace class can be too broad, by which I mean that it generates more information than you need for the probelm at hand. This brings us to the IFCID.

Each trace class is associated with specific trace events known as Instrumentation Facility Component Identifier (IFCIDs), pronounced “if-kid.” An IFCID defines a record that represents a trace event. IFCIDs are the single smallest unit of tracing that can be invoked
by Db2.

In some cases, it can make sense to avoid activating trace classes altogether and start traces specifying only the IFCIDs needed. This way, you can reduce the overhead associated with tracing by recording only the trace events needed.

You can use trace classes 30 through 32 for starting IFCID-only traces. These classes have no predefined IFCIDs and are available for a location to use. Consider the following -START TRACE command, which starts only IFCIDs 1, 2, 42, 43, 107, and 153:

-START TRACE(PERFM) DEST(GTF) CLASS(30) IFCID(1,2,42,43,107,153)

If you do not specify the IFCID option, only those IFCIDs contained in the activated trace classes are started. The maximum number of IFCIDs per trace is 156.

Because this task can be tedious, if you decide to trace only at the IFCID level, it is probably a good idea to use a performance monitor that starts these IFCID-level traces based on menu choices. For example, if you choose to trace the elapsed time of Db2 utility jobs, the monitor or tool would have a menu option for this, initiating the correct IFCID traces (for example, IFCIDs 023 through 025).

There are several hundred different IFCIDs. Most IFCIDs contain data fields that describe events pertinent to the event being traced. Some IFCIDs have no data; instead they merely mark a specific time. For example, IFCID 074, which marks the beginning of Terminate Thread, has no data field.

Certain trace events of extended durations require a pair of IFCIDs: one for the beginning of the event and another for the end. These pairs enable the computation of elapsed times. Other trace events that are not as lengthy require only a single IFCID. 

You can find the IFCIDs associated with each trace class in the IBM DB2 Command Reference in the section describing the -START TRACE command. But that manual does not describe the purpose of each IFCID. A list describing each IFCID can be found in the SDSNIVPD(DSNWMSGS) data set. It also contains DDL and LOAD utility statements that can be used to create DB2 tables containing IFCID information that can be easily queried by SQL.

For example, the following query retrieves field descriptions for performance trace records for particular classes:

       ownerid.TRACE_TYPES         B
AND    CLASS IN (1,2,3,4,5,6,7)

You can modify the query as needed to retrieve only the specific IFCID information you need.


This series of blog posts has offered up a high-level introduction to Db2 tracing and what type of information is available using traces. A comprehensive strategy for monitoring the performance of your Db2 subsystems and applications is an important part of establishing a best practices approach to database performance management. Use this information to help you on your journey to an efficiently monitored Db2 for z/OS environment.

Of course, this information is just the beginning. Keep checking back on this blog for additional useful information, and/or make sure you have a copy of my book, Db2 Developer's Guide, which contains over 1,600 pages of Db2 guidance and advice.

Wednesday, June 13, 2018

Db2 for z/OS Performance Traces Part 2 - Global, Monitor, Performance, and Statistics

In Part 1 of the series on Db2 performance traces we provided a general overview, as well as a discussion of the Accounting and Audit trace classes. Today, in Part 2, we will discuss the remaining 4 trace classes: Global, Monitor, Performance, and Statistics.

Global Trace

The global trace is one I hope you never have to use. It produces information that is used to service Db2, so you'd only start a global trace at the direction of IBM if you are having some sort of trouble. A global trace records information regarding entries and exits from internal Db2 modules as well as other information about Db2 internals. 

Global trace records are not accessible through normal tools that monitor Db2 performance. Most sites will never need to use the DB2 global trace. You should avoid it unless an IBM representative requests that your shop initiate it.

A global trace can add significant CPU overhead to your Db2 subsystem.

Monitor Trace

Quite a bit of useful performance monitoring information is recorded by the Db2 monitor trace. Most of the information in a monitor trace is also provided by other types of Db2 traces. The primary reason for the existence of the monitor trace type is to enable you to write application programs that provide online monitoring of Db2 performance.

Information provided by the monitor trace includes Db2 statistics and accounting trace information, as well as details of current SQL statements.

There are ten groups of DB2 monitor trace classes:
  • Class 1: Standard accounting data
  • Class 2: Entry or exit from DB2 events
  • Class 3: DB2 wait for I/O or locks
  • Class 4: Installation-defined monitor trace record
  • Class 5: Time spent processing IFI requests
  • Class 6: Changes to tables created with DATA CAPTURE CHANGES
  • Class 7: Entry or exit from event signaling package accounting
  • Class 8: Wait time for a package
  • Class 9: Statement level accounting
  • Class 10: Package detail
  • Class 11 through 28: Reserved
  • Class 29: Dynamic statement detail
  • Class 30 through 32: Local use

The overhead that results from the monitor trace depends on how it is used at your site. If it is used as recommended, class 1 is always active with 2 and 3 started and stopped as required, the overhead will likely be minimal but will depend on the activity of the Db2 system and the number of times that the other classes are started and stopped. If you make use of the reserved classes (30 through 32), or additional classes (as some vendors do), your site will incur additional overhead.

Do not start the monitor trace using DSNZPARMs unless online performance monitors in your shop explicitly require you to do so. It is best to start only monitor trace class 1 and to use a performance monitor that starts and stops the other monitor trace classes as required.

Some online performance monitoring tools do not use the monitor trace; instead, they read the information directly from the Db2 control blocks. Sampling Db2 control blocks requires less overhead than a monitor trace but it can be disruptive if the tools encounters bugs.

Performance Trace

The Db2 performance trace records an abundance of information about all types of Db2 events. You should use it only after you have exhausted all other avenues of monitoring and tuning because it consumes a great deal of system resources. When a difficult problem persists, the performance trace can provide valuable information, including SQL statement text, a complete trace of the execution of SQL statements, including details of all events, all index accesses and all data access due to referential constraints.

There are 22 groups of Db2 performance trace classes:
  • Class 1: Background events
  • Class 2: Subsystem events
  • Class 3: SQL events
  • Class 4: Reads to and writes from buffer pools and the EDM pool
  • Class 5: Writes to log or archive log
  • Class 6: Summary lock information
  • Class 7: Detailed lock information
  • Class 8: Data scanning detail
  • Class 9: Sort detail
  • Class 10: Detail on BIND, commands, and utilities
  • Class 11: Execution unit switch and latch contentions
  • Class 12: Storage manager
  • Class 13: Edit and validation exits
  • Class 14: Entry from, and exit to an application
  • Class 15: Installation-defined performance trace record
  • Class 16: Distributed processing
  • Class 17: Claim and drain information
  • Class 18: Event-based console messages
  • Class 19: Data set open and close activity
  • Class 20: Data sharing coherency summary
  • Class 21: Data sharing coherency detail
  • Class 22: Authorization exit parameters
  • Class 23 through 29: Reserved
  • Class 30 through 32: Local use

When all Db2 performance trace classes are active, you will experience significant overhead, perhaps as much as 100% CPU overhead by each program being traced. The actual overhead might be greater (or lesser) depending on actual system activity. The ­overhead when using only classes 1, 2, and 3, however, typically ranges between 5% and 30%.

Performance traces must be explicitly started with the -START TRACE command. Starting the performance trace only for the plan (or plans) you want to monitor by using the PLAN parameter of the -START TRACE command is wise. Here’s an example:


Failure to start the trace at the plan level can result in the trace being started for all plans, which causes undue overhead on all DB2 plans that execute while the trace is active.

Furthermore, due to the large number of trace records cut by the Db2 performance trace, system-wide (Db2 and non-Db2) performance might suffer because of possible SMF or GTF contention. 

Statistics Trace

The final type of Db2 trace is the statistics trace, which contains information pertaining to the entire Db2 subsystem. This type of information is particularly useful for measuring the activity and response of Db2 as a whole. Information on the utilization and status of the buffer pools, DB2 locking, DB2 logging, and DB2 storage is accumulated by the statistics trace.

There are ten groups of DB2 statistics trace classes:
  • Class 1: Statistics data
  • Class 2: Installation-defined statistics record
  • Class 3: Data on deadlocks, lock escalation, group buffers, data set extension, long-running units of recovery, and active log shortage
  • Class 4: Exceptional conditions
  • Class 5: Data sharing statistics
  • Class 6: Storage usage
  • Class 7: DRDA location statistics
  • Class 8: Data set I/O
  • Class 9 through 29: Reserved
  • Class 30 through 32: Local use

The estimated overhead of the statistics trace is low. Approximately 1% to 2% CPU overhead per transaction is incurred by the Statistics trace.

Db2 cuts a statistics trace record periodically based on the setting of the STATIME subsystem parameter (DSNZPARM). STATIME is specified as a time interval, in minutes, and can range from 1 to 60 minutes. It is a good practice to set STATIME to 1, thereby specifying 1,440 statistics intervals per day. The information accumulated by cutting these statistics trace records can provide valuable details for solving complex system problem. 
By analyzing the evolutional trend of statistics, sometimes the cause of problems can become evident that would otherwise be difficult to track down. 

Even though 1,440 records sound large, in reality the amount of data collected is small when compared to the typical volume of accounting trace data collected. An additional thousand or so SMF records should not cause any problems, while at the same time offering valuable system information.

Next time...

This concludes the overview of the types of Db2 tracing that are available. In part 3, we will examine where trace records can be written as well as more narrow tracing using IFCIDs.

Friday, June 08, 2018

Db2 for z/OS Performance Traces Part 1 - Intro, Accounting, and Audit

When it comes to Db2 for z/OS performance management, one of the first things you need to understand is the traces that are available that must be started in order to cause Db2 to track performance metrics. You can think of a Db2 trace as a window into the performance characteristics of aspects of your workload. Db2 traces record diagnostic information describing specific performance events. As Db2 operates, and workload is run against it, Db2 writes trace information based on the traces types that have been started. The information is written out in such a way that it can be read and analyzed by DBAs and performance analysts.

There are six types of Db2 traces available, each one pertaining to a different category of Db2 performance and operations. These six types of traces are:

  • Accounting - performance information about the execution of DB2 application programs 
  • Audit - information about Db2 DDL, security, ­utilities, and data modification activities
  • Global - information for servicing Db2 
  • Monitor - data useful for online monitoring of the Db2 subsystem and application programs 
  • Performance - detailed data about Db2 events, enabling database and performance analysts to pinpoint the causes of performance problems 
  • Statistics - information regarding the Db2 subsystem’s use of resources 

There are two ways to start most Db2 traces: you can specify the appropriate DSNZPARMs to start traces when Db2 starts up or you can issue the -START TRACE command to initiate specific traces when Db2 is already running. 

Each trace is broken down further into classes, each of which provides information about aspects of that trace. 

Accounting Trace

The accounting trace is one of the most important traces for judging the performance of Db2 application programs. Accounting trace records can contain information regarding:
  • CPU and elapsed time of the program
  • Usage of the EDM pools
  • Locks and GETPAGE page requests, by buffer pool, issued by the program
  • Number of synchronous writes
  • Thread wait times
  • Type of SQL issued by the program
  • Number of COMMITs and ABORTs issued by the program
  • Program’s use of sequential prefetch and other DB2 performance features (RLF, distributed processing, and so on)

There are ten groups, or classes, of Db2 accounting traces:
  • Class 1: Standard accounting information
  • Class 2: Entry or exit from DB2 events
  • Class 3: Elapsed wait time in DB2
  • Class 4: Installation-defined accounting record
  • Class 5: Time spent processing IFI requests
  • Class 6: Reserved
  • Class 7: Entry or exit from event signaling package accounting
  • Class 8: Wait time for a package
  • Class 10: Package detail
  • Class 11 through 28: Reserved
  • Class 30 through 32: Local use

Most organizations start Db2 accounting classes 1, 2, and 3 together because they provide a great deal of information with only a low overhead. Furthermore, you cannot run class 2 or 3 without also running class 1.

Accounting trace classes 7 and 8 provide performance trace information at the package level. For class 7 accounting, trace information is gathered for packages executed, and for class 8 information, the amount of time an agent was suspended in Db2 for each executed package is collected. If you have already activated accounting trace classes 2 and 3, adding trace classes 7 and 8 incurs only a minimal additional cost.

Audit Trace

The audit trace is useful for installations that must meticulously track specific types of Db2 events, particularly to support regulatory and compliance requirements. Not every shop needs the audit trace. However, those wanting to audit by AUTHID, specific table accesses, and other Db2 events mayfind the audit trace invaluable. 

Of course, as of Db2 10 for z/OS, an alternate approach to auditing with audit policies is available. This requires use of SECADM, which is not pervasively implemented, though.

There are eight categories of audit information that can be provided using the audit trace:
  • All instances in which an authorization failure occurs, for example, if USER1 attempts to SELECT information from a table for which he or she has not been granted the appropriate authority
  • All executions of the DB2 data control language GRANT and REVOKE statements
  • Every DDL statement issued for specific tables created by specifying AUDIT CHANGES or AUDIT ALL
  • The first DELETE, INSERT, or UPDATE for an audited table
  • The first SELECT for only the tables created specifying AUDIT ALL
  • DML statements encountered by DB2 when binding
  • All AUTHID changes resulting from execution of the SET CURRENT SQLID statement
  • All execution of IBM DB2 utilities

Again, there are groups, or classes of trace information that can be started. In the case of the audit trace, there are eleven trace classes:
  • Class 1: Attempted access denied due to lack of authority
  • Class 2: GRANT and REVOKE 
  • Class 3: CREATE, ALTER, and DROP statements against audited tables
  • Class 4: First change made to an audited object
  • Class 5: First read made against an audited object
  • Class 6: BIND information for SQL statements on audited objects
  • Class 7: Assignment or change of an AUTHID
  • Class 8: Utility execution
  • Class 9: Installation-defined audit trace record
  • Class 10: Trusted Context information
  • Class 11: Audit administrative authorities
  • Class 12 through 29: Reserved
  • Class 30 through 32: Local use

The overhead of the audit trace depends greatly on the amount of data that is being audited. Things like transaction frequency, modification frequency, DDL activity, and so on will dictate the actual amount of overhead required.

Next Time...

So far, we have only looked at two of the six types of Db2 traces. Be sure to check back later for Parts 2 and 3 of this blog series that will introduce the remaining types of traces and additional useful information on dealing with Db2 performance tracing.

Monday, May 21, 2018

The Db2 12 for z/OS Blog Series - Part 22: Function Levels 501 and 502 (Continuous Delivery)

If you have heard anything about Db2 Version 12 chances are that you have heard about continuous delivery. Instead of waiting 2 to 3 years for a new version of Db2 to be released, new functionality will be continuously delivered on a regular basis. The idea is to bring Db2 into the modern age of development practices where releases are small and quick, instead of large and slow.

So instead of waiting for the next version, Db2 professionals now wait on new Function Levels, where a Function Level identifies a set of new enhancements that can be enabled in Db2 for z/OS.

Of course, this means that a lot of internal practices and procedures had to be re-engineered and established at IBM, so there have not been many new Function Levels since Db2 12 was first released back in October 2016. There was Function Level 501 in early 2017, which basically added a simple new built-in function, LISTAGG.

The LISTAGG built-in function produces a list of all values in a group. An optional separator argument can delimit items in the result list. For example, specifying a comma as the separator produces a comma-separated list. An optional ordering can also be specified for the items within the group. So for example:

             AS EMPLOYEES

This will return a comma-separated list of employee last names by department number.

Unless you needed the capability of LISTAGG in your applications there was no reason to migrate to Function Level 501. Except, of course, to test out moving to a new Function Level, which is the primary reason that IBM released LISTAGG as a Function Level. And that was it until recently…

Function Level 502 (FL502) was made available by IBM in late April 2018. This is the first “real” Function Level with multiple new capabilities that may entice your shop to implement it. 

Here are the capabilities introduced in FL502:

The first new feature bolsters DFSMS data set encryption (which is part of the Pervasive Encryption for IBM Z solution introduced with the z14). With FL502 we get KEYLABEL management capability for z/OS DFSMS data set encryption. You can manage the key labels for z/OS DFSMS data set encryption to transparently encrypt Db2 data sets. 

DFSMS can be used to encrypt various types of Db2 data sets including Db2-managed table space and index space data sets, data sets that are used by Db2 utilities, and sequential input and output data sets. 

After moving to FL502 an administrator (DBA, security admin, system admin or storage admin depending on your shop) can enable z/OS DFSMS data set encryption for your Db2 data sets.

Additionally, IBM offers a free tool, IBM z Systems Batch Network Analyzer (zBNA), which can be used to help estimate the costs of DFSMS data set encryption for your Db2 data sets. Additionally, the Db2 Statitistics Trace has been enhanced to report CPU time, which you can look at to help determine which data sets to encrypt.

The second enhancement enabled with FL502 is the ability to cast an explicit numeric value to a graphic string value. All of the numeric data types are supported. So you can use the GRAPHIC or VARGRAPHIC built-in functions and/or the CAST specification to cast numeric values to graphic string values. Regardless of whether CAST or GRAPHIC/VARGAPHIC functions are used, the result is Unicode (UTF-16), and the context must support Unicode data.

Implementing Function Level 502

You can activate Function Level 502 from Function Level 501, 500, 100, or as part of migration from Db2 11 (with z/OSMF only). Function level 502 requires catalog level 502, and tailoring the catalog for level 502 requires function level 500 or 501. Take care before activating any new Function level by making sure that you understand what Function Levels are, how they are delivered, and the current state of your Db2 subsystems.

You can easily view the current state of your Db2 subsystems by using the -DISPLAY GROUP command. It will show you the current Function Level, the high Function Level ever activated (which might be higher than current if you fell back), and the highest possible Function Level (based on the APARs that have been applied to your Db2 system).

What's Next?

Things are likely to speed up in terms of new Function Levels for Db2. Now that IBM has had time to implement new internal development proceudres and get them all test out appropriately, we should start seeing new capabilities more frequently than once a year... perhaps as frequently as quarterly. So make sure that you are ready to review every new Function Level as it is made available and make plans to activate the ones that deliver functionality that you need.

Another thing to keep in mind is that Function Levels are cumulative. So you cannot implment say, Function Level 502, without also getting the capabiltiies of all previous Function Levels (in this case, just 501). So be prepared and understand what activating a new Function Level means!

Welcome to the new world of continuous delivery in Db2 for z/OS… and take a look at how the new capabilities in Function Levels 501 and 502 might be useful at your shop and to your applications.

Monday, May 14, 2018

Improving Performance by Caching Select Statement Results

A significant trend in the database world these days is moving more data management and processing into memory. If you can bypass disk I/O you can achieve tremendous performance gains. There are many reasons for this, but of course, the most important reason is that disk access is much slower than memory access.

It is orders of magnitude more efficient to access data from computer memory than it is to read it from disk. Memory access is usually measured in microseconds, whereas disk access is measured in milliseconds (1 millisecond equals 1000 microseconds).

So, it makes sense to process as much data as we can in memory. This is the reason that most popular RDBMS use buffer pools to cache data in memory. But buffer pools are not always sufficient for caching everything that is needed. Just take a look at IBM Db2 12 for z/OS and the index FTB feature, which caches unique index structures in memory – not in any buffer pool.

FTBs, or Fast Traverse Blocks, provide in-memory index optimization. FTBs are stored in a separate area outside of your buffer pools. You must allocate additional real memory for it in Db2 12. This memory is an optimized structure for fast index lookups that provide random index access. The FTB contains the non-leaf structure of the index.

QuickSelect for Db2

Of course, FTBs are just one technique to optimize your memory usage. Another technique you might consider is to deploy QuickSelect for Db2, an in-memory optimization product from Log-On Software (an Israeli company that has been providing tools and services to the mainframe market for decades). QuickSelect for Db2 caches query results in memory. This is a different concept than buffer pools which cache table and index space pages or in-memory tables which cache entire tables.

By caching the results of frequently run queries in memory with QuickSelect for Db2, you can optimize the SQL statements that recur frequently in your environment. QuickSelect runs as a started task and it automatically caches repetitive query results in self-managed memory above the bar (64-bit).

QuickSelect returns the same data DB2 would – but from its cache. QuickSelect will identify and cache only frequently used data via a threshold parameter. This saves both CPU and I/O, and therefore can reduce elapsed time and improve application response times. All other requests are satisfied as usual by Db2.

Which leads to the question: which queries should be cached? QuickSelect helps you to identify ideal queries with its Survey mode which automatically finds SQL queries suitable for caching.

The results cached by QuickSelect are the actual data retrieved by Db2 queries. So the next time that same query runs (with the same host variable values) the results are quickly returned from the QuickSelect cache instead of invoking process-heavy disk I/O operations. Think of all the repetitive SQL that you are running in your shop every day – and all of the redundant I/O that is used to read that data over and over again.

But, you may be asking, what happens if the data changes? Well, QuickSelect is aware of all changes that are made, whether by a Db2 utility like Load, Reorg and Recover, or by your applications issuing Insert, Update, Delete, and Truncate statements. Data changes on the tables that are cached are detected by QuickSelect in real-time using various techniques. QuickSelect invalidates the cached data for the updated tables automatically, thus data integrity is preserved. Further queries will cause QuickSelect to cache the updated data, thereby continuing to save. QuickSelect runs in a multiple LPAR environment and supports update sensitivity by using XCF functions.

In other words, QuickSelect returns the same answer as Db2 does… only faster. And, of course, that is the most important thing.

If you were paying attention as you read through this so far you will have noticed that there are no application changes required whatsoever to enable QuickSelect. And you do not have to Rebind your programs to take advantage of QuickSelect, either!

What happens if you shut down QuickSelect? There should be no impact on your application programs; processing reverts entirely back to Db2. The data will no longer be returned from the QuickSelect cache, instead it will again be retrieved by Db2 just like before you used QuickSelect.

QuickSelect: The Latest Features

Log-On Software is not resting on its laurels; new features are added to QuickSelect on a regular basis. For example, in Febrauary 2018 auto-configuration was added to QuickSelect, thereby easing configuration in a Data Sharing environment.

With auto-configuration the QuickSelect group configuration is dynamically determined. It is no longer required to pre-configure QuickSelect groups. QuickSelect now automatically detects the active members in the Db2 group, if any, and automatically establishes a connection between all QuickSelect servers that correlate to the current Db2 configuration. This is much simpler than the previous requirement of hard-coding QuickSelect grouping using pre-defined parameters.  

The Bottom Line

QuickSelect for Db2 offers a different approach to in-memory caching by focusing on query results. QuickSelect for Db2 customers have been able to significantly reduce CPU during peak times – the times set by the rolling four-hour average that impact your monthly IBM software bill.

For example, one customer, a large European commercial Bank was able to turn off an entire CPU after installing QuickSelect in production due to the CPU savings they achieved. In this case, the customer was saving more than 10 billion SQL statements during a typical 12-hour window.

The bottom line is that in-memory processing can save CPU, improve performance and save money. Wise organizations will look into multiple ways of exploiting memory to achieve results, including potentially looking at novel software solutions like QuickSelect for Db2.