By John Pickett
When your applications and IT environment are the driving force behind your company's growth, the reliability and availability of your servers becomes increasingly important. By now, I'm sure you are familiar with the often discussed stats around the impact to businesses in the event of a disaster:
- 93% of companies that lost their data center during a disaster for 10 days filed for bankruptcy in 1 year
- ½ the businesses filed for bankruptcy immediately without data management over the same time period. (U.S. National Archives and Records Administration)
Once upon a time, starting in the late 60's, the mainframe was called upon to provide high levels of availability to protect the company's crown jewels and intellectual property. There of course was a premium price for the availability and reliability achieved by the mainframe.
That was then - this is now. Since 2000, the HP Superdome has been a cornerstone of companies' data centers for the same reasons - to provide a high volume of transactions with while achieving very high levels of reliability and availability. Yeah, but the mainframe started doing it first, so it must be better - right? Is first better? Don't know - let's go ask the first portable computer company about this. Oh wait, you can't - Osborne Computer is no longer in business. I know it's an offbeat example, but the point is - first does not mean better.
So what are some of the components designed in to the HP Superdome that helps provide the high levels of availability and reliability? A partial list is below:
- Intel® Itanium® 2 processors
- Intel Cache Safe Technology
- Electrical isolation of hard partitions
- Dynamic processor resiliency
- Double chip spare
- Dynamic memory resiliency
- Dynamic multi-pathed I/O
- Advanced I/O error recovery
- PCI computer bus online addition, replacement, and deletion
- Link-level retry and link retraining
- Spare channels
- Clock redundancy
- Power redundancy
- Redundant and hot-plug fans and power supplies
- Dynamic Root Disk
- Online addition, replacement, or deletion of I/O
- Dynamic partitions
But what does it all mean? One of the key building blocks for the HP Superdome is the Cell Board (processors, main memory, cell controller, voltage regulators, data buses & links to I/O slots) and has a Mean Time Between Failure (MTBF) of 57 years! This MTBF is based on actual field replacement rates. Needless to say the mainframe is not the only system to deliver the reliability and availability needed for business critical applications.
The real difference is when you take in to account the 3 year TCO cost difference between the z10 mainframe and an HP Superdome. The HP Superdome has a TCO that is 1/8 that of the latest z10 mainframe. That difference would be even greater had we used older System z mainframes for our comparison.
You find this white paper and other resources at HP's Migrate from IBM Mainframe to HP site, and you can also take your own mainframe TCO Challenge .
