Traditionally, providing disaster recovery and high availability for Power Systems running IBM i has relied on three main technologies, clustering such as PowerHA SystemMirror for i, SAN based replication such as EMC SRDF or IBM Storwize MetroMirror/GlobalMirror, or logical replication solutions such as Mimix, iCluster, MAXAVA, QuickEDD, etc. Now IBM adds a new flavour called Geographically Dispersed Resiliency (GDR).
GDR takes some technology from both PowerHA and LPM (Live Partition Mobility) to provide a DR solution with an aggressive RPO and a less aggressive RTO. The RPO can be close to zero, as GDR uses GlobalMirror/MetroMirror or the EMC SRDF equivalent to transmit changes from the production SAN array to the DR SAN array. The less aggressive RTO is because GDR has to wind down the VM or IBM i LPAR on production then do what is in essence an abnormal IPL form the SAN on the DR system. The LPM technology was the ability to easily move a live LPAR including all of it’s users and currently running jobs to another LPAR.
Here is how it all hangs together:
The chart above indicates that the management of GDR is done through PowerVC consoles by issuing AIX command in an AIX LPAR called K-SYS (for Control System – seems IBM is not great at spelling). Apparently the switch is as simple as a command which tells GDR that “I want to switch to System 2” and GDR does the rest – bringing down production and IPL’ing the backup. You will also noticed from the chart no mention of IBM i since this technology has been available for AIX for a couple of years and to GDR, IBM i is just another VM. The downside for IBM i shops is there will be a learning curve for AIX commands.
IBM is not targeting this towards the large shops, who have DS8000s and are probably already invested in PowerHA, but to smaller IBM i shops running the popular Storwize arrays. As of June 23, the SVC storage server – which is not nearly as popular among IBM i shops – is now supported with GDR, along with IBM i, Storwize, and DS8000 arrays. They are also targeting cost savings as since the customer software is not actually active on multiple machines at the same time and the recovery site is a cold vs warm backup, savings could be made on licensing. Also the product is licensed as a single LPAR (the AIX K-SYS) not by the core, which could also provide costs savings over traditional logical replication solutions.
The good news is that there are a number of technologies that will provide DR for Power Systems running IBM i, the bad news is the same as the good news! How do you choose which technology is right for your business? In some cases it’s a single solution, in others it may be a mix of hardware, software and services. Mid-Range has experts who have forgotten more about HA and DR on multiple platforms than most shops will ever know. We can help you implement a DR solution or re-examine one that you already own, and we can host it for you too. Mid-Range can help.