Big Blue And ITG Build Case For IBM i Resiliency

December 8, 2014 Dan Burger

Business resiliency and business continuity are terms that are mistakenly used as synonyms with disaster recovery. A disaster is the worst type of disruption to business continuity. It is also the most infrequent. Like death from a thousand paper cuts, a variety of little things can play havoc with business resiliency. How much havoc is related to the value a company attaches to any aspect of the business that does not run at peak performance.

Downtime is the enemy of business continuity. Every year the tolerance for downtime edges closer to zero. For some shops zero tolerance has been in effect for years. The financial loss is too extreme regardless of duration. For the most part, these are large businesses running big iron and they live in fear of losing their business continuity. They work to minimize it and understand the lack of continuity extends beyond localized systems to effect enterprise-wide systems and in many instances no customers and partners go unscathed.

To demonstrate the importance of resilient systems, IBM hired the research and analytics firm ITG. IBM pitted its IBM i 7.2 operating system on Power 7+ hardware with PowerHA SystemMirror for i clusters compared to one system running MicrosoftWindows 2012 servers with SQL Server 2014 AlwaysOn clusters on Intel E5- and E7-based hardware and another system consisting of Windows 2012 servers running Oracle Exadata X3-8 and X4-2 systems with Real Application Clusters (RAC) and Data Guard.

The focus on this comparison (the fairness of the comparison can be argued separately), comes down to two important issues: cost of downtime and risk exposure.

The three-year costs of downtime by platform shows Microsoft Windows Server (green bar) to be the most costly and the IBM i/Power Systems (blue bar) to be the least costly in the financial services category.

This report is based on large enterprise customers in two groups: financial services companies and supply chain companies. The financial services companies included a bank, an insurance company, and a services company. The supply chain companies included a manufacturer, a retail chain, and an industrial distributor. Revenues at these companies varied between $1 billion and $12 billion. Employee counts were between 2,000 and 50,000.

In the supply chain category the IBM i/Power Systems combo (blue bar) is far less expensive than the Microsoft Windows Server (green bar) and Oracle Exadata (orange bar) systems.

The cost of downtime for the supply chain companies was steepest for the manufacturer. That expense was $1,305 per hour. Downtime costs for the retail chain was $833 per hour and the distributor was losing $685 per hour when down. The factors affecting these numbers were supply chain disruption, lost sales, and customer penalties and remedial costs.

Downtime costs for the financial services companies were highest for the bank. It was figured at $320 per hour. The insurance company downtime costs were figured at $164 per hour and the services company downtime was pegged at $149 per hour. The totals were attributed to customer attrition, lost fee income, and other costs.

According to ITG, business costs attributable to outages were 8.3 times higher for the Windows/SQL Server system and 3.4 times higher for the Oracle Exadata system when compared to the IBM i system. The calculation took into account outages of less than three hours duration.

Although the direct application of this comparison may apply to only a few large IBM i shops, everyone in the IBM i community should be aware of this study because it represents one of the few times IBM puts some effort into extolling the virtues of the platform. Business continuity is important for the small business, too.