The following tip is excerpted from Chapter 1, Choosing your server, from of our expert e-book, "Windows servers and storage." This chapter touches on aspects of server hardware, beginning with architectures--systems components and interconnects -- and ways to build servers from these components.
Types of server architectures
It makes economic sense for manufacturers to offer composable systems, which can be configured to meet a broad spectrum of needs by plugging together the right subsystems. An important consideration when selecting a server is the amount of computing or processing power needed. Because there are strict limitations on computer performance available from a single processor, a key dimension in composability is the number of processors deployed in a system.
There are several ways in which multiple processors may be deployed within a system. They include the following:
- Symmetric Multiprocessing (SMP): This method arranges processors so each sees all system memory and all I/O, allowing programs to run on any processor (or processor core) and access all system resources.
- Clustering: This technique builds the system by connecting separate, independent computer subsystems, in which each has its own processor, memory, storage and I/O.
- Grid computing: A sort of clustering variant, grid computing employs the clustering approach using real, separate computers as its building blocks, and a LAN or WAN as the system.
Each of these three approaches has its own strengths and weaknesses, making each appropriate for different classes of applications.
As a side note, not all servers are Intel x86 (IA32) based. As you will see in market figures later in this section, there is a minority share of the market owned by processor architectures like Sun Microsystems, Inc., SPARC, IBM POWER and Intel Corp. Itanium. While the processor architectures vary, the underlying physics remain the same for all machines. You should look at the architecture as part of the decision process, but it should not be the deciding factor per se. A better approach is to consider the applications needed to run the business and choose the appropriate system (price, performance, life-cycle cost, etc.).
About the authors:
René J Chevance is an independent consultant. He formerly worked as chief Scientist of Bull, a European-based global IT supplier.
Pete Wilson is Chief Scientist of Kiva Design, a small consultancy and research company specializing in issues surrounding the move to multi-core computing platforms, with special emphasis on the embedded space. Prior to that, he spent seven years at Motorola/Freescale.