In our last newsletter we discussed the overall movement to redesign the typical data center LAN in general and the movement to reduce the number of layers in the design in particular. In this newsletter we will continue that discussion and will include how the traditional assumptions and LAN design guidelines are giving way to a new set of assumptions and guidelines.
The traditional data center LAN has been built around a number of assumptions that have stood up to the test of time – at least up until now. One assumption was that nearly all of the traffic within the data center was client/server related. This type of traffic is often referred to as “north-south” traffic. A part of this assumption was that the relatively small amount of traffic that went between servers was readily visible on physical switch ports. This type of traffic is often referred to as “east-west” traffic. While this assumption may still be true for some IT organizations, many IT organizations have told us that they have seen a great growth in east-west traffic within their data centers.
One of the traditional LAN design guidelines was that each server was dedicated to a single application and had a single IP address and a single MAC address. This assumption made total sense in a world of physical servers and it makes no sense in a world of virtualized servers. Analogously, in a traditional data center it makes sense to assume that the physical separation of servers satisfies most of the requirements for security and for the delineation of departmental resources and that the static nature of the application environment means that it is perfectly acceptable to have manual processes for server deployment or redeployment. These assumptions are of very questionable validity in a highly virtualized data center.
In our next newsletter we will discuss an RFI for data center LAN switching.
