Availability and scalability for large businesses

Large businesses choose Db2 for z/OS because they need a robust database server that ensures superior availability and scalability.

You might be thinking that the terms enterprise server and mainframe imply that very large businesses use a product like Db2 for z/OS.

You might ask the question: Why do large businesses choose Db2 for z/OS? The answer is, Because these companies need a robust database server that ensures superior availability and scalability.

Superior availability and scalability in a Parallel Sysplex® environment are the key features that distinguish Db2 for z/OS from other database servers. Because of these qualities, Db2 for z/OS is widely deployed in industries that include:

  • Major credit card companies
  • Banks
  • Insurance companies
  • Brokerage companies
  • Credit information companies

These are companies that process very high volumes of transactions that require millions of concurrent updates every day.

Consider some examples.

  • The volume of trading that goes on at the major stock exchanges can reach over 1,000,000,000 shares in a single day.
  • A brokerage company might have a network of thousands of financial advisors and millions of customers who need online access to highly sensitive financial information daily.
  • A transportation company might deliver more than 10 million packages in a single day. Each package requires several steps in the delivery process, such as pick up, transit points, and final delivery. The status of the package can be shown to customers on the web.
  • A credit information company needs to provide millions of credit reports each day, while keeping the data current with more than 100 million updates in a single day.
  • A sports website provides statistics, results, and live updates for their viewers. The site must continue to provide this information to their viewers quickly and efficiently during hours of peak demand.

You can easily understand why these businesses need the database system that processes these transactions to be continuously available, scalable, and secure. These enterprise systems must be available to customers who are searching for and relying on their services 24 hours a day.

  • Systems must provide continuous availability.

    If you are waiting for a financial transaction to process and the application that runs that transaction suddenly fails, you might lose the opportunity to make a stock trade at a critical time. The key objective of high availability is to ensure that a system has no single point of failure.

  • Systems must be scalable.

    As businesses grow, their data processing needs also grow. Business ventures, such as mergers, acquisitions, and new services, or new government regulations, can accelerate how quickly the data processing needs of the business grow. As rapid growth occurs, companies need a way to scale their business successfully.

    Companies need a large database system that is designed to easily absorb ongoing additions of new types of information and application processes without sacrificing performance or availability. That database system should never impose a constraint on growth. As businesses add more computing capacity, the database system must expand accordingly to ensure that businesses gain the full advantage of the added capacity and have continuous access to their data.

  • Systems must be secure.

    The needs for protection and regulatory compliance are expanding greatly. Customers must be able to trust the database when they use it to manage valuable information such as finances or personal information. IBM zSystems has a long history of system integrity and security. Db2 security responds to the needs.

The following scenarios describe how a large international bank benefits from these Db2 for z/OS strengths to provide the highest quality of service to its customers.

Scenario 1: merging unrelated applications

Bank mergers occur often. As two banks combine operations, how does the newly formed bank merge unrelated applications?

Db2 for z/OS data sharing in a Parallel Sysplex environment provides the solution that the new bank needs so that the two banking systems can be merged.

Parallel Sysplex clustering technology in Db2 is the answer to availability and scalability. A Parallel Sysplex is a cluster, or complex, of z/OS® systems that work together to handle multiple transactions and applications. This technology implements a data sharing design.

The Db2 data sharing design gives businesses the ability to add new Db2 subsystems into a data sharing group, or cluster, as the need arises and without disruption. As applications run on more than one Db2 subsystem, they can read from and write to the same set of shared data concurrently.

The Parallel Sysplex can grow incrementally without sacrificing performance. Parallel Sysplex architecture is designed to integrate up to 32 systems in one cluster. In a shared-disk cluster, each system is a member of the cluster and has access to shared data.

An integral component of a Parallel Sysplex is the coupling facility, a mechanism that coordinates transactions between the different members within a cluster. Other solutions attempt to implement similar capabilities through software, but messaging by using software can cause high overhead and directly impact the ability to scale and perform.

When Parallel Sysplex technology is used, the applications from each bank can easily be integrated into a data sharing group and can access shared data.

Scenario 2: balancing varied workload peaks

The bank runs batch jobs every night and the online workload is running close to 24 hours a day. How can the bank run varied workloads, keep them balanced, and avoid problems at peak times?

Db2 works closely with the z/OS Workload Manager (WLM) component. WLM provides the best way to run mixed workloads concurrently, and data sharing gives the bank a lot of flexibility in how to run the workloads.

Parallel Sysplex technology is designed to handle varied and unpredictable workloads efficiently. The Workload Manager ensures that the bank's workloads are optimally balanced across the systems in the Sysplex.

For example, when the bank adds a new subsystem or the workload becomes unbalanced, data does not need to be redistributed. The new subsystem has the same direct access to the data as all existing subsystems in the data sharing group.

Data sharing works with WLM to give the bank the flexibility it needs to handle peak loads easily. WLM provides the ability to start up servers and subsystems on demand, based on predefined service goals. For example, the bank can start data sharing members to handle peak loads at quarter-end processing, and stop them when the quarter-end peak finishes.

Db2 is the only data server on IBM zSystems to take full advantage of WLM capabilities.

Scenario 3: 24-hour online banking

The bank creates a website to provide online banking to its customers 24 hours a day. Now the DBMS can never be out of service for maintenance activities. How can the bank apply maintenance to its DBMS if it needs to be operational 24 hours a day?

Data sharing and Parallel Sysplex technology give the bank a way to apply software maintenance (a planned outage) while always keeping a subset of its Db2 subsystems up and running.

The Parallel Sysplex environment provides multiple paths to data and builds redundancy into the coupling facility to avoid single points of failure. With Parallel Sysplex technology, the bank can add maintenance to one member at a time while their systems continue running and remain up-to-date on service. The technology also allows the bank to migrate to a new software release by applying the new release to one member at a time. With this design, the bank avoids outages.

In the event of an application or a system failure on one system (an unplanned outage), the Workload Manager ensures that other systems within the sysplex can take over the full workload. Again, the bank avoids outages.