The benefit of having them all in the same box is that a 10GB internal, switched network can be used to link them all together. This improves throughput. Another advantage of a single housing is that if a problem occurs, or just when an internal server architecture is being developed, all of the systems specialists will be able to work in the same room, which rarely happens in real life. This should speed things up from many angles.
The brain behind the system is an uber controller, the Unified Resource Manager (URM). Not only does this control the mainframe, switches and blades but it also monitors operations. If an element fails, or looks like it is failing, the system will automatically flag the problem, call in the systems manager and order a spare from IBM.
To allow the URM to do its work means that all systems have to be verified. This is why the system can only use IBM blades and even then only ones that have been through the validation process.
"The zEnterprise can offer up to a 90 per cent saving in space," said Barrie Heptonstall, mainframe and Power systems sales leader at IBM. "It's taken three years to develop and if customers think of it as a Cloud in a box, it wouldn't be far from the mark."
The systems will be available from September and the pricing structure has yet to be revealed.
