edf40wrjww2CF_PaperMaster:Desc
One of the ways in which organizations with large scale server deployments (consisting of hundreds to tens of thousands of servers, such as those found in enterprise data centers) are dramatically cutting costs is to replace large servers with highly compact rack-mountable forms. These smaller form factors dispense with costly individually-attached hardware (such as power supplies and device interfaces) in favor of shared resources among the servers in a rack.
The densest of these forms is the blade server, which began shipping in 2002. In addition to reducing the hardware, electrical and square footage costs, blade servers are more manageable than traditional servers, since, in addition to being hot pluggable and simplifying cable configurations (and incorrect configurations can be a major source of downtime), a single server can be used to manage all other servers in the same shared chassis. While some blade servers have internal disks, they tend to have lower performance and capacity than SCSI disks, a fact which is helping to drive adoption of diskless blade servers used in combination with storage networks (NAS and SAN).
The development of diskless blade servers introduced a challenge for versions of the Windows operating system prior to the Windows Server 2003 server release, since Windows boot procedures were originally developed with the requirements that the boot disk be directly attached to the server, and that the operating system have access to the boot volume at all times.
With the release of Windows Server 2003 (and updates to Windows 2000), the Windows platform supports boot from SAN capabilities, without requiring that the dedicated boot disk be local to the server. These capa ...