Use IBMDB2pureScaleFeature and QReplication for scalability and business

To calmly cope with today's increasingly globalized and competitive markets, such a data processing architecture is indispensable. This architecture can flexibly grow as the future strategic needs grow, ensures business continuity in the event of component faults, maintenance activities, and disaster events. For some enterprises, even an hourly shutdown may cause millions of dollars in revenue loss, let alone

To calmly cope with today's increasingly globalized and competitive markets, such a data processing architecture is indispensable. This architecture can flexibly grow as the future strategic needs grow, ensures business continuity in the event of component faults, maintenance activities, and disaster events. For some enterprises, even an hourly shutdown may cause millions of dollars in revenue loss, let alone

To calmly cope with today's increasingly globalized and competitive markets, such a data processing architecture is indispensable. This architecture can flexibly grow as the future strategic needs grow, ensures business continuity in the event of component faults, maintenance activities, and disaster events. For some enterprises, even an hourly shutdown may result in millions of dollars in revenue loss, not to mention damages to the company's reputation and potential loss of customers. Global enterprises run across different time zones to provide business services all the time. Offline windows retained for system maintenance and upgrade no longer exist. Distributed enterprises must be able to provide similar services in each geographic location, and avoid network faults or transmission time. This article describes an architecture to address these availability requirements.

In December 2009, IBM launched the DB2 pureScale Feature for Enterprise Server Edition. The DB2 pureScale Feature is built on sophisticated design features from IBM DB2 for z/OS software. DB2 pureScale Feature is designed to meet the needs of many customers and provides:

-Almost unlimited capacity: You can easily expand your system by adding more machines to the cluster.
-Application transparency: You can use your existing applications without any restrictions.
-Continuous availability of a single site: provides a "full activity" architecture with internal redundancy.
-Lower total cost of ownership (TCO): by supporting simplified deployment and management of advanced technologies, the total cost of ownership is reduced.

Since the first version was released in 2004, Q Replication technology has provided low-latency and large-capacity Replication functions for DB2. Q Replication supplements the features of pureScale and provides the following features:

-To defend against disk and site faults, copy the database to a remote site.
-The continuous availability during the upgrade and full site maintenance period allows the application to be transmitted to another site before the maintenance is complete, and the sites can be re-synchronized after the completion.
-Server Load balancer unload and Real-Time Report: Detach the report application to another database to eliminate any possible contention for business-critical workloads and support real-time data analysis and reporting.
-Defends against data corruption and maintains a secondary database copy. Compared with the primary database copy, the changes are delayed and can be quickly recovered from user or application errors.

By extending the DB2 pureScale Feature with Q Replication, you can achieve scalability, reliability, and transparency, while ensuring the protection and continuous availability of remote replicas. This article will discuss the considerations of this solution and introduce how to deploy the Q Replication and DB2 pureScale Feature in combination.

DB2 pureScale Feature utilizes a shared Disk Database Based on the DB2 for z/OS data sharing architecture. It introduces mature technologies of DB2 database software from the mainframe into open systems. Using DB2 pureScale Feature brings the following important benefits:

• Almost unlimited capacity: DB2 pureScale Feature provides almost unlimited capacity by supporting on-demand addition and deletion of DB2 members. DB2 pureScale Feature can be expanded to 128 members and has a highly effective centralized management tool that supports a variety of different scaling features. DB2 pureScale Feature also utilizes a technology called Remote Direct Memory Access (RDMA), which provides a very effective inter-node communication mechanism, simplifies the extended Feature of DB2 pureScale Feature.

• Application transparency: an application that uses the DB2 pureScale database does not need to understand any knowledge or worry about data of different members in the cluster. DB2 pureScale Feature automatically routes the application to the members it deems most appropriate. DB2 pureScale Feature also provides native support for many syntaxes used by other database vendors so that these applications can run in the DB2 pureScale environment with minimal or even no changes required. In many cases, the benefits of DB2 pureScale Feature are achieved without modifying the application.

• Single-site continuous availability: DB2 pureScale Feature provides an active-Active configuration. Therefore, if a member fails, the processing operation can continue on the remaining active members. During a member's fault, the transaction data being processed on the faulty member is temporarily unavailable until the database is fully restored, which usually takes several seconds.

• Lower TCO: DB2 pureScale Feature helps reduce TCO through its integrated, simplified deployment and maintenance features. The DB2 pureScale interface handles deployment and maintenance of components integrated in the DB2 pureScale Feature.

Since the first version was released in 2004, Q Replication has provided high-performance Replication for DB2 on Linux, UNIX, and Windows and DB2 on z/OS, including DB2 z/OS data sharing, these copies are based on Log Capture and transaction latency. Q Replication can replicate a large number of changes to thousands of DB2 tables across thousands of kilometers, often with a latency of less than one second. Q Replication uses WebSphere MQ for effective data transmission and temporary storage of Replication changes. Q Replication supplements the features of pureScale and provides the following features:

-Defense against disk or site faults: the DB2 pureScale instance provides almost unlimited scalability and Defense Against Faults of each member. However, traditionally, only independent Disk redundancy arrays (RAID) are configured) use a single data copy on the disk. Copying a database to a remote site provides additional defense against disk array faults and a site disaster solution. When Q Replication is used, there is no restriction on the distance between sites, and the recovery site is allowed to be far away from the main site, so as to avoid the same risk 1 as the main business site. Because the Replication process using Q Replication is almost real-Time, database Recovery is almost real-Time, and the restoration Time Objective (RTO) is only a few seconds.

-Continuous availability during upgrade and maintenance: Planned Maintenance and migration take into account service interruptions rather than disasters. By transmitting applications to another website that uses Q Replication synchronization, you can achieve business continuity during system maintenance and updates. Upgrade examples include changing hardware, operating systems, DB2 versions and applications, and any data center maintenance that may require system shutdown.

-Workload unload and Real-Time Report: In the DB2 pureScale instance, the application may be analyzed and reported to compete with the locks of online business events, affecting database performance. By transferring a report application to another database, you can generate reports and perform analysis based on real-time data without affecting key business applications. Q: The Replication process is asynchronous and does not affect the application response time. Q Replication can replicate data between systems with huge differences, or even perform data conversion. For example, the main site may be a DB2 pureScale instance containing 16 members, using Q Replication to bring a subset of the database to near real-time (often with an average latency of several seconds) copy to another pureScale instance with fewer members, or even copy it to a non-DB2 database. Q Replication provides the ability to copy a subset of a database or even a subset of database transactions (for example, you can only copy selected operations or exclude transactions executed by a specific user ).

-Avoid data corruption through a delayed copy: for restoration at a time point, Q Replication can maintain a secondary copy of the database, which is a little longer than the primary database copy. You can use this secondary copy to recover data from user or application errors in the primary database.

Secondary databases can also be used to query reports at time points. The Q Replication technology can be used to easily store delayed copies of databases, because changes captured from the source database are accumulated in a WebSphere MQ receiving queue on the target database. The Q Apply program on the target system can use an applydelay parameter to run 2. The program will continuously Apply these changes after the specified number of seconds after the change is submitted on the source database. You can also use the applyupto parameter to Apply changes in batches. In this case, the Q Apply program applies the changes made before the predefined time point and then stops.

In this way, changes are saved to the target to defend against disasters, achieving the potential recovery point objective (RPO) of sub-second in seconds, even if the data is not immediately applied. Because changes are captured and transmitted to the target queue in near real time, DB2 logs on the source can be archived normally; they do not need to be retained for replication. If failover is required, the recovery time target (RTO) is the time required for the replication function to clear the backlog of changes in the receiving queue on the target. The Q Apply program can quickly clear the backlog of changes, often clearing millions of cumulative changes in minutes, thanks to its parallel application technology.