It’s fair to say that cloud has dramatically changed the landscape for IBM i workloads. While electing to err on the side of caution, most businesses start by migrating development or test partitions to the cloud. But as more production workloads are migrated should you adapt your approach to HA/DR when considering the move?
IBM i logical replication built on remote journaling is the foundation of Maxava’s HA suite and has been implemented over two thousand times around the globe. Despite an ever-changing landscape with hardware replication now commonplace and with the advent of the cloud, it continues to be the mainstay of high availability. For this reason, logical replication can be described as having the Lindy Effect – the longer something exists, the longer it will exist, in the main due to its flexibility to adapt to different eras and the variety of environments that we now see. As well as being a regular choice for cloud migrations, Maxava’s HA suite runs as efficiently in the cloud as on-premise, with both HA and DR options available. To help minimize the risk of having all your resources with one cloud provider it’s good practice to use multiple providers to host your SOURCE and TARGET partitions, or at the very least to utilize two or more from the growing list of regions these providers have built. Maxava’s HA suite supports both scenarios.
Most public clouds provide a 4-hour SLA (Service Level Agreement), which relates to the amount of time the physical IBM Power System which your partition is hosted by, will be recovered within. It is a hardware SLA and is normally underpinned by a similar agreement between the cloud provider and IBM for hardware, plus SWMA (IBM software maintenance agreement) for operating system and licensed program support. In addition, providers normally offer uptime assurances along with financial compensation for times when uptime falls below the agreed figure.
In the event of a hardware failure, what the 4-hour SLA doesn’t cover is the time it takes to recover your partition. You will still need to consider your application and the state it was in at the time of the failure, along with elements such as access path rebuilding and the recovery of any damaged objects. It’s worth noting that during any outage you will be unable to transact, and any prolonged outage has the potential to harm business reputation.
One of the neat advantages that cloud provides is having the ability to start your partition on an alternate node, but be sure to remember 3rd party licensing if you are considering this feature as many IBM i software keys are linked to a serial number which will change as you move nodes. To alleviate these challenges cloud providers often allow partition pinning to provide the guarantee that your partition will always be started on the same node (serial number). When this feature is offered cloud providers are unable to sell this resource to other customers and as such you are charged whether the partition is active or inactive. It also detracts from one of the advantages of running workloads in the cloud.
A cloud migration initiative provides you with the ability to re-evaluate your HA/DR requirements, and your first step should always be working out your RTO (Recovery Time Objective) and RPO (Recovery Point Objective). In layman’s terms RTO is the targeted duration of time within which a business process must be restored following a disruption or disaster to avoid unacceptable consequences associated with a break in business continuity. RPO is the maximum targeted period in which data might be lost from an IT service due to a major incident. Having both the RTO and RPO figures at hand is imperative when designing the architecture along with supporting services and software for any HA/DR deployment.
Maxava have long worked with MSPs (Managed Service Providers) providing solutions for their Power Clouds, whether this is HA between a pair of partitions hosted by a MSP, or as part of a DRaaS (Disaster Recovery as a Service) offering and extend these solutions to public cloud providers such as Skytap and the IBM Cloud as well as IBM’s Power Virtual Server.
Comments