MACH 11 - Tell me more about it
MACH 11 stands for Multinode Active Cluster for High Availability.
It is an extremly flexible High Availability Technology which is fully integrated in the Informix DBMS Kernel and builds upon the efficient and nearly linear scalable multithreaded architecture.
CLR - Continuous Log Restore
IDS 11.1 supports the concept of continuous log restore. A secondary instance is placed in rollforward mode and logical logs from the primary instance are constantly applied. onbar as well as ontape support the continuous restore of logical logs thru the new command switches (-C and -X). This could be done in delayed mode to prevent the immediate replication of logical errors from the primary node.
Previous Informix releases did already support the concept of mirroring a complete instance to a secondary machine for many years. The secondary instance could be used for read access (distributing queries and reports to relieve the primary instance) and could also act as a failover instance. This concept is called HDR (High Availability Data Replication) and the existing functionality it is still available. The traffic between the HDR Primary and Secondary node can be encrypted with IDS 11.1.
The HDR concept has been significantly extended by MACH 11. It is now possible to mirror the primary instance to any number of secondary nodes which could be distributed worldwide at different places. Those mirror nodes - also named RSS (Remote Standalone Secondary) nodes - store the data locally and allow read access to the data. Resource intensive queries and reports could be moved from the primary instance to those secondary instances.
Due to the new fully duplexed communication protocol implemented in IDS 11.1 those RSS nodes can be distributed worldwide even on slower lines. The full network traffic can be encrypted.
In addition to the mirroring of a primary instace, MACH 11 alow allows the usage of SDS (Shared Disk Secondary) nodes similar to the Oracle RAC concept. The SDS nodes run on separate machines, have their own bufferpools, but do not store a separate copy of the data. They share the disks with the primary node (for example in the scope of a SAN environment). A new SDS node can be created within a few minutes. You just need to add an entry to the sqlhosts file, copy over the onconfig and change a few parameters. After that you execute oninit and the new SDS node is up and running - just as easy as it can be. A failover from the primary instance to a SDS node can be accomplished in a few seconds.
MACH 11 is an extremly flexible high availability technology that allows any combination of the following functionalities:
CLR (Continuous Log Restore)
Ability to keep a second IDS node in constant rollfoward mode, applying the logical logs from the primary node
HDR (High Availability Data Replication)
1:1 Primary-/Secondary-Replication with automatic failover and read acces on the secondary node
RSS (Remote Standalone Secondary)
1:n Primary-/Secondary-Replication with read acces on all secondary nodes
SDS (Shared Disk Secondary)
1:n Primary-/Secondary-Configuration with shared access to disks, failover support und read access on the secondary nodes
A combination of MACH 11 and Informix ER (Enterprise Replication) is also smoothly possible. ER allows the asynchronous replication of individual tables to an arbitrary number of other IDS nodes and offers the possibility of read/write to the data on any node (Update Anywhere) if desired.
MACH 11 is the consistent continuation of the superb replication technology that Informix Dynamic Server offers since many years. With the flexibility of MACH 11 IDS will stay far ahead of DB2 when it comes to high availability and will even stay ahead of Oracle because MACH 11 is better integrated and much easier to configure and maintain than a comparable Oracle technology.
IBM Looks to 'Cheetah' to Speed Up Blade Servers
Availability Solutions with IDS 11
Build your own MACH Cluster
The IDS Dev. Edition Wiki - Building an IDS MACH 11 Cluster