Chapter 2: Hardware Concepts

---

2.1 Introduction

In the previous chapter, we started with a basic introduction to this book and discussed the modern business requirements covering the various abilities that a system should possess, like availability, scalability, manageability, etc. Then we stated that clustered solutions are required to meet most of these requirements. However, to determine if a clustered solution was really the option, we stated that a complete feasibility study followed by a capacity planning is required.

Meeting these requirements ensures that the computer systems that are used today and in the future should be robust enough to handle the dynamic growth of users, both locally within the organization and across the world where users could potentially access any system via the Internet.

This requires that the systems selected to support the various enterprise systems should be able to satisfy the business requirements listed in the previous chapter. While the basic requirements are possible, to maximize the potential to support those requirements depends on the underlying technology that is selected. This stipulates another requirement that, while selecting the technology to run the system, for example, the hardware platform, the hardware technology, storage systems, database systems, etc., considerable care and study should be provided before arriving at a suitable choice.

In this chapter the various hardware architectures such as SMP, MPP, NUMA, etc., will be discussed at length. We will also explore the usage of clustered SMP architecture as a prudent solution for businesses with data processing needs. This architecture allows businesses to provide data storage and retrieval to users and to support increased growth of users through linear scalability. We will discuss areas where many systems are joined together to work as a cohesive unit, how they support not only the scalability factor but also the availability factor. When one system or node in the cluster fails, the users have other nodes that could be used to route their request to get to the required information.

How do these systems meet the scalability requirement? In this chapter we will discuss clustering in relation to the hardware architectures and discuss how clustering will help meet the scalability requirement. Finally, we will also look at the storage subsystems. Selection of the right storage systems helps in meeting the availability requirement. For this, it is important to select the appropriate storage architecture while taking into consideration the performance requirements. The various RAID technologies and the appropriate benefits of using one over the other will be discussed.

---