IP Storage Networking: Straight to the Core

Storage networks are fueling a rapidly adapting architectural model for storage. Today's storage infrastructure continually adapts to keep pace with application deployment and data creation. Innovative applications and the sheer amount of information stored shift processing bottlenecks throughout the system. Bottlenecks often arise when systems are architected based on the rich experience of yesterday 's requirements with imperfect knowledge of how those requirements will evolve over time. Both entrenched storage vendors and enterprising upstarts aim to alleviate processing bottlenecks through the use of pioneering technologies deployed as storage system control points.

The simple split between servers and storage, or initiators and targets, implies a need for network connectivity. The introduction of a storage network allows previously restricted storage functions to operate across this split to provide an end-to-end solution. A complete understanding of these storage functions, potential areas for deployment, and best practices for optimized solutions allow administrators to properly architect storage infrastructures .

Chapter 4 begins by outlining the basic storage services required to build a robust, resilient data storage repository. By starting with the basic storage functions, we keep our fulfillment objectives in mind rather than simply talking about the control points abstractly. Classified as storage service deliverables, these functions include capacity, performance, availability, recovery, security, and manageability.

A look at the historical evolution of storage helps clarify where storage functions and associated intelligence have resided in the past and potential areas where they will reside in the future. Although the hype for new technology may cause us to believe that we face a completely new set of circumstances, technology transformations follow historical patterns, many of which apply to storage. Examples in distributed systems and routing apply to the evolution currently taking place in storage networking. Understanding the history of storage evolution and its technology cycles helps us understand today's challenging environment and prepare us for future decision making.

Ranging from the simplest to most complex storage networks, functions and intelligence reside within three basic system areas ”host, fabric, and subsystem. Looking at specific storage functions, we examine the pros and cons of where they reside and why. The dramatic changes underway come from the ability to place intelligent storage services in the fabric. Fabric-based solutions bring fluidity to storage network control points.

Networking plays a crucial role as the link between hosts and subsystems. Brokering all storage transactions, traditional interconnects historically set finite boundaries on storage control points. Today, the convergence between I/O technologies such as Fibre Channel and networking technologies such as Ethernet shatter previous conventions of storage design. With an ever-increasing amount of data looking for a means to get from A to B, and a corresponding requirement to heavily invest in storage fabrics , choice of storage interconnects can dramatically affect total cost of ownership and return on investment.