DB2 DBA, how to explain the business value of DB2

Many technicians can easily discuss the details of DB2 technology and confidently talk about query parallelism, data compression, WebSphere MQ integration, large object management, JDBC and ado.net drivers, data sharing on mainframe Parallel Sysplex, DB2 for Lin UX, Unix, and Windows (LUW) multidimensional clusters, and so on. But what happens if the person who talks is a member of the management? The main concerns of corporate managers are revenue growth, cost control, product quality, and the time the product is put into the market. Generally speaking, these people do not care about the lock granularity, server memory management and SQL statement optimization such technical issues. They do not care about the characteristics of DB2 technology itself (although DB2 technology is cool), but rather about what role DB2 can play in achieving organizational goals. This article will help people using DB2 to discuss DB2 technology from the perspective of business value.

Many technicians can easily discuss the details of DB2 technology and confidently talk about query parallelism, data compression, WebSphere MQ integration, large object management, JDBC and ado.net drivers, data sharing on mainframe Parallel Sysplex, DB2 for Lin UX, Unix, and Windows (LUW) multidimensional clusters, and so on. But what happens if the person who talks is a member of the management? The main concerns of corporate managers are revenue growth, cost control, product quality, and the time the product is put into the market. Generally speaking, these people do not care about the lock granularity, server memory management and SQL statement optimization such technical issues. They do not care about the characteristics of DB2 technology itself (although DB2 technology is cool), but rather about what role DB2 can play in achieving organizational goals. Anyone who uses DB2 every day should be able to discuss DB2 technology from the perspective of business value.

Through my experience in CheckFree Corp., I've summed up a list of key areas in which DB2 technology can provide the value that business people can feel.

Scalability vs. as business grows

Single-server DB2 systems, whether running on mainframes or on high-end Linux, Unix, or Windows servers, can provide significant throughput for OLTP, business Intelligence (BI), or composite workloads. Throughput is largely due to the fact that DB2 leverages 64-bit server memory addressing, novel I/O optimization features (such as List prefetching), Pre optimized SQL (DB2 professionals call it static SQL), and advanced workload management capabilities. However, in a rapidly expanding business environment, the amount of data access requests can grow rapidly, and the ability of a single-server system may not be sufficient to handle future requests. Business leaders certainly do not want the growth of the business to be limited by the scalability of the data server. This is an important part of scale expansion (scale-out), and scalability is the area where DB2 really takes advantage.

In this context, the term "scale extension" refers to the ability to spread the workload for a single logical image database across multiple physical servers. There are two DB2 scaling solutions: Data sharing on a mainframe cluster (called Parallel Sysplex), and the implementation of database partitioning Feature (DPF) on a Linux, Unix, or Windows server cluster. Both of these technologies are the leading technology in the industry. DB2 for z/OS data sharing enables enterprises to perform concurrent read/write access to a shared database from up to 32 DB2 subsystems that can run on many mainframe systems or on a small number of physical servers with multiple DB2 subsystems on each server. This solution is not the only scalable solution for shared data DBMS on the market, but no other technology can provide DB2 for z/OS data sharing groups such as good CPU efficiency (very low CPU overhead for read/write data access between nodes that are really concurrent).

DB2 with DPF capabilities provide unmatched scale scalability in Linux, Unix, or Windows environments. Hundreds of servers can be configured in a DB2 DPF system; Each server provides access to a physical subset of a single logical image database (a hashing algorithm disperses rows from a given database table to the node specified by the DBA). There are other unshared (shared-nothing) DBMS scaling solutions in the marketplace, but other solutions are not as easy and flexible as DB2 with DPF, because the DPF feature is embedded in the DB2 for LUW data service engine.

People are still arguing about the pros and cons of shared data and unshared DBMS architectures, but both of these DB2 solutions are appropriate for the underlying server platform. The CPU overhead of DB2 for z/OS data sharing is very low because the functions it uses are distributed in an optimized manner throughout the Parallel Sysplex software architecture: z/OS operating system, DB2 DBMS, coupling facility control Cod E (it manages the shared memory structure used for global locks and data buffering) as well as the CICS transaction manager or the DB2 Connect distributed client Gateway (if these components are in the configuration). This optimization is possible because DB2 for z/OS data sharing requires only one operating system and one chipset (IBM System z microprocessor). Such a feature distribution is not possible in the DB2 for LUW environment, as such environments need to support multiple operating systems and multiple server hardware platforms; Therefore, the DB2 for LUW scale expansion solution is based on the best unshared clustering technology. Either way, the organization will have the desired effect: the DBMS does not hinder business growth.

Efficiency vs. lower total cost of ownership

When assessing the costs associated with various data service solutions, there is often a concern about the cost of obtaining hardware and software licenses. The price of software and hardware is important, but this is a small percentage of total cost of ownership (TCO) associated with the DBMS. Other factors that affect costs include:

The number of people required to administer the database system;

The efficiency of using hardware resources (CPU and hard disk storage);

(a) training costs for technology;

The difficulty of making different database systems work together in an enterprise;

First, DB2 for z/OS, because it replaces very expensive mainframe based solutions in some areas. The following factors can affect cost control on the System Z platform:

Scale economy

On DB2 for z/OS systems, a very large workload can be handled, and the data Service mainframe can run smoothly even if it is more than 90% utilization over a few hours. As transaction volumes grow, the cost of a single transaction for a platform is significantly reduced.

Price/Performance Trend

Although the System Z platform is a fairly expensive one (because it provides cutting-edge hardware and software technology), the cost of unit computing capacity (typically measured in millions of instructions per second or MIPS) has declined over the past few years. The price of mainframe software, whether from IBM or other vendors, is even more competitive than it used to be.

Manageability

Organizations can handle very large workloads on mainframe DB2 systems without the need for a large number of support personnel. One reason for the astonishing productivity of DB2 for z/OS systems Programmers and DBAs is that many companies offer rich mainframe DB2 tools, and another benefit is that DB2 for z/OS produces rich trace data that the aforementioned tools can format. And the cost is often very low. DB2 for z/OS support staff can also benefit from certain platform features, such as system-managed storage, which allows the z/OS operating system to place tables and indexed datasets in the hard disk subsystem (the larger the database, the more significant the efficiency advantage of the system-managed storage).