IBM i High Availability Technology & Hardware Overview

February 16, 2016 Disaster Recovery Blog, High Availability Blog Articles, Legacy System Blog Articles
In association with : Vision Solutions


Making the right decision for your company’s high availability (HA) solution for IBM i environments requires a clear understanding of its architecture and technologies. There are many HA solutions to choose from, and the right, informed decision can mean the difference between your company missing or meeting its recovery time objective (RTO) and recovery point objective (RPO) in the event of a disaster.

The Basics

On the IBM i platform there are two main ways to maintain HA during a system outage: logical replication and hardware HA. Each solution offers individual and unique strategies to ensure business continuity and the safety of your data.

Logical Replication

Logical replication maintains a hot backup copy of the IT environment that is available to the business at the time of failover from its recovery systems—it is remote journal-based data resiliency. This means that every change to the data is logged and sent to your company’s IT system where it is accessible to be applied. The operating system (OS) journal function ensures that recorded data is reflective of its most recent state and is securely written to the backup system—this helps your company achieve the best possible RPO. Your RTO will depend on how quickly the HA applies the journaled transactions to your production server from the target system.

Hardware HA

Hardware HA is either a disk-based replication that works at the TotalStorage SAN level, or a page-replication solution at the OS level. With both configurations, data that is written to a defined set of drives is duplicated in an additional set of drives. When the memory page is written to the disk with a hardware-HA configuration, the first SAN makes a sector-by-sector copy of the changes on the second SAN. IBM i treats memory and disk storage as one, and so the data being copied is encapsulated and treated as a subset of the system address range.

Hardware HA requires switchable independent auxiliary storage pool (IASP) technology. This is where local journaling of all objects takes places to support recovery after a failure. IASP is a collection of disk units that can be brought online or taken offline, and be kept separate from the rest of the data on a storage system connected to IBM i. This process mitigates the risk of disk failure because they designate part of the disk unit for various object storage.

The Factors

Hardware HA and logical replication are very different choices, and each comes with specific advantages and limitations. Recovering from a disaster without damaging existing data and within your company’s minimum RTO and RPO, requires careful attention. Your company must consider how the different technologies will impact your:

  • Bandwidth between systems
  • Number of copies of the data
  • Timely save of data in memory
  • Journaling protection for system recovery
  • Latency with synchronous communication
  • IASP vary-on and recovery from system outage
  • Target system availability and HA protection
  • Unplanned downtime behaviors

Along with the above areas, your company must consider how they will maintain data synchronization, share SANs between platforms, and address the system ASP and the basic user ASP (SYSBAS) replication needs.

The Decision

Each company’s mission and operations is unique, and so must be their HA environment. A system that is operational 24/7/365 is optimal, but to achieve this a company must properly prepare to meet their RTOs and RPOs. In order to have guaranteed retention, it is imperative that your company has a disaster recovery (DR) plan in place.

Logical corruption can defeat your HA solutions, which includes software glitches and bugs in your applications, or human error. This is why you need to have a DR plan working in tandem with your HA solution. HA solutions are tasked with making your data available and redundant, but they don’t care about the state of your data. So, your data can be highly available, but if it is logically corrupted, it’s useless. On the other hand, physical corruption is corruption at the IO subsystem level, where Structured Query Language (SQL) servers ask the OS to write data to disk. The request to write the data can be mangled, and it will leave you with tainted data.

HA is important, and no matter what strategy you choose to utilize, you will always need a DR plan as well. To define what your company’s best options are for HA and DR solutions, contact Baseline Data Services today at (317) 707-3941and our highly trained engineers will ensure your data and applications will be safe, no matter the disaster.

Back to blog list