Most businesses begin to look at cloud, at least initially, for disaster recovery or high availability, and utilize IBM i logical replication tools such as the Maxava HA Suite to help reduce the risk and downtime associated with achieving this. Others look to the cloud seeking development and test partitions. Furthermore, those workloads which have variable demands on resource or that have the fewest regulatory restrictions can also be considered good candidates for cloud. Relatively speaking, IBM i was somewhat late to the cloud party and what started with a trickle several years ago has now turned into a steady stream of businesses exploring cloud options. Some of this interest can be attributed to the coronavirus pandemic with businesses keen to close data centers having struggled to get staff into these sites to perform manual tasks such as changing backup tapes. Options now are plentiful with public cloud options at IBM, Skytap and Google providing Platform-as-a-Service (PaaS), added to the more traditional Managed Service Provider Power Cloud Infrastructure-as-a-Service (IaaS) offerings.
Public cloud providers allow the choice of operating system, which is normally in-line with current the IBM supported versions, namely V7R3 or V7R4. Private clouds operated by Managed Service Providers often have the capability to offer older versions such as V7R1 and V7R2 too. Both providers offer a good selection of pre-installed licensed programs with additional options available at a further charge.
Once you’ve selected a cloud provider your next and probably biggest hurdle is weighing up the best approach to migrating your applications and data to the cloud. Tape is not an option as cloud-based partitions have no concept of physically attached devices. The data
involved with IBM i applications can be sizeable and your options depend on two factors: how much data you have, and how large an outage your business can afford. If you’re moving a workload such as development or test where downtime is not really a concern a backup and restore-based approach may well be a viable option. But IBM i environments across the globe run applications in banking, finance, manufacturing, healthcare, distribution, logistics and telecommunications, and any outage, even a planned one, is unlikely to be welcomed by the business. Outages increase the risk of prolonged downtime along with a potential impact to revenue, productivity in addition to reputational damage.
With BRMS the de facto backup solution of many IBM i shops, it can certainly be utilized to aid your journey to the cloud, although as previously stated, not with physical tape media. When BRMS is paired with the chargeable IBM Cloud Storage Solutions for i (5733-IC1), virtual media can be used, although it’s worth noting that due to the nature of virtual media, with saves writing to virtual volumes on disk, you will require the same amount of available storage as the data you are backing up. For many freeing up sufficient disk space can be a real challenge. With BRMS and IBM Cloud Solutions for i, the virtual volumes can be automatically FTP’d to a target of your choice, including IBM Cloud Object Storage, IBM Spectrum Protect, Amazon Web Services or Google Cloud Storage. These locations can be used as an interim location before the data is retrieved and then restored to your cloud partition. Cloud providers may also be able to provide you with an FTP or SFTP server for use as an alternative interim location. IBM Cloud Storage Solutions for i is really aimed at SMB environments where the amount of data being saved doesn’t exceed 2TB.
For larger environments looking to move more data to the cloud there are hardware devices available, such as the Cloud Mass Data Migration (IBM), or the Data Box (Skytap on Azure). These devices are physically shipped to you, you point your backups to them before they are dispatched back to the cloud provider who will make them available on your partition of choice for restore purposes. Of course, any transactional changes to data in-between when backup and restore into your cloud partition are not captured. This window can extend to several days due to manual transportation of these hardware devices, and as a result are best suited to the initial seeding of the target, priming it ready for logical replication.
Logical replication for IBM i, the cornerstone of the Maxava HA Suite, is built on remote journaling, itself integrated into the operating system. It is used to keep data and the IFS synchronized in near real-time and is also a popular choice in IBM i shops for HA (High Availability) or DR (Disaster Recovery). The Maxava HA Suite can be used to replicate data to an empty target partition, although by firstly seeding the target you eliminate the need to transfer everything over the network and thus seeding can speed up cloud migrations. When seeding the target cloud partition Maxava can be configured to synchronize data from a specified starting point, normally in line with when the full save was performed, it will then keep the data and all transactions up to date throughout your migration, providing a valuable safety net as well as allowing you to test in parallel, eliminating the risk normally associated with migrations. Once you’re satisfied that the data is synchronized, that 3rd party licensing has been addressed and that communication are in place, a role swap is conducted to make the cloud partition live.
In summary, unless you have a relatively small amount of data, and a huge window that will allow you to save, transfer and restore your data you’re going to require a combination of data transfer (either via the network or by physical device), and a logical replication tool like the Maxava HA Suite.
After completing your cloud migration, you can continue to adopt the Maxava HA Suite in-between on-premise and cloud partitions, or even for added resiliency from one cloud provider or region to another.
This article is written by Ash Giddings, Product Manager at Maxava