IBM Database Informix 11 flagship availability new architecture

IBM's latest database software, Informix Dynamic Server (IDS) 11, strengthens database service provisioning, data provisioning, and load-splitting mechanisms, improving the resilience and performance of database sizing.

The main change in Informix Dynamic Server (IDS) 11 is the enhanced availability of both highly available replication (HDR) and increased support for RSS and SDS two server functions, which reinforces the stability of the database while balancing the effects of data provisioning and load sharing.

In the past, an IDs HDR used two servers as an instant backup. One as the primary server, another is a secondary server, different from other database systems in the use of snapshot backup, IDs in the main server's data have a change, only a small number of SQL log records, passed to the secondary server, secondary server, according to the record execution of SQL instructions, synchronized adjustment of the contents of the database , consistent with the primary server. Primary server downtime, as a local backup secondary server, can be converted into a primary server to continue to provide services.

In version 11, IDs extends HDR technology, adding RSS servers that can span regions, as well as remote database synchronization through SQL log records, to achieve offsite backup effects.

Further, on the local side, the SDS server that shares the disk with the HDR primary server is also added. SDS is used as a service backup for primary servers, unlike RSS data backup. For users, SDS provides the same new query and change functionality as the primary server, but when the user's SQL instructions enter the SDS, a switch layer is used to determine the SQL instruction type, and the SQL required to change the database is all directed to the primary server, which is called a write redirect. The primary server controls all the differences to the database, while SDS shares the burden of the query. Because of shared disks, no data synchronization is required for all actions.

In order to disperse traffic, the general database system uses cluster design. However, once multiple servers are encountered to write to the database, it is necessary to control the order of writes, one at a time, and the other to stop the read action to ensure the consistency of the data. Moreover, the more distributed servers, the more write conflicts may result, and the longer the database locks up, thus causing the bottleneck of the database.

Cooper Information Technology director Lin Yin said: "IDS 11 new architecture just solves this problem." "writes data to the primary server, avoids write conflicts, provides only secondary servers for queries, and provides continuous service." And he added: "Such a framework is very suitable for the blade server, when the load reaches the peak, you can increase the number of server slices to divert." ”

To speed up the enterprise upgrade IDs version, IBM offers a variety of scenarios, with the exception of a few versions requiring a phased upgrade, most versions can be directly upgraded to version 11.