Taking into account a comprehensive analysis of the existing or proposed storage infrastructure, implementation becomes the critical factor. A big factor in implementation is the acquisition of upgrade hardware and software or a completely new solution. Coupled with acquisition of the storage networking components is the scheduled installation, testing, and transfer of production storage data and workload I/O processing. This is followed by the initialization of the systems management cycle of activities.
Acquisition of storage capacity is an exercise in commodity purchases driven in most data centers by price. The best price per MB usually gets the business and drives storage vendors to differentiate their offerings in the firmware and software solutions that become part of the system. However, the storage networking business has turned this upside-down, with the necessity of new networking devices and the enhanced storage arrays and tape systems that must be FC-enabled. This has placed price in a more balanced perspective with other major purchase factors, such as reliability, service, and quality.
Consequently, the following guidelines can be used to assist in data-center acquisition strategies when purchasing storage networking solutions.
Competitive Bids Its important to gather at least three competitive bids for the configuration you are seeking. Ensure that the requirements for storage networking, be it SAN or multiple NAS solutions, are available to the vendors you have decided to work with. Be advised that total solution offerings that are provided through storage vendors and systems integrators allow them an additional revenue source.
OEM Knowledge Most storage networking vendors, especially larger system vendors, provide their solutions as a composite of external components supplied by the third-party companies. These companies provide their component, be it hardware or software, as an Original Equipment Manufacturer (OEM) supplier, and place their name on the device. This is commonplace in storage networking with most SAN solutions being made up of OEM suppliers. This is not a bad thing, but its important to compare apples to apples when considering competitive bids. In other words, dont pay extra for the same FC equipment, such as switches, HBAs, or routers that is available through a competitive solution.
Storage vs. the Network Try to separate the storage part from the network part. This is especially helpful in SAN acquisition, given that the network portion is likely to expand at a different rate than the actual storage arrays. The storage capacity plan should provide for incremental upgrades through the planning cycle, generally a year in length. During this time, storage array acquisition can be part of the competitive bidding process, but should be negotiated with as little bid lock-in as possible. In other words, have the vendors consider bidding on the entire plan (for example, a complete year of requirements) to obtain a better price, service, and vendor commitment.
Many of the examples shown in the book depict storage networks supporting three common types of workloads: OLTP, Web Internet Based, and Data Warehouse. More details on these I/O workloads can be found throughout the book with additional guidelines on identification and estimating in Chapter 17. Typically, SAN configurations are comprised of combinations of 8-, 16-, and 32-port FC switches, with disk arrays commensurate with storage capacities that have been estimated with workloads. Within NAS installations, typical configurations can encompass both departmental solutions and support within the data center for more sophisticated applications. Another important point to consider is the external factors that influence the installation. These are Ethernet network modifications within NAS installations and should be considered prior to installation. SAN configurations are subject to the inclusion of intersystem -link ports (ISLs) and an integrated FC-SCSI bridge into a tape library.
It is important to define, configure, and install a permanent test installation environment. Putting a small configuration in place provides essential first-case experiences in the configuration and operation of both SAN and NAS installations. This also provides a test bed for testing future software and hardware upgrades while enabling an application testing facility.
Use the test installation to initiate a pseudo-management practice. Management becomes a very challenging activity, especially when operating the new devices and complexities in the SAN. It also is the most rapidly evolving practice, with constant change occurring in software tools and accepted practices. Additional discussion of storage network management topics can be found in Part VI.
Develop a production turnover activity where a formal change window is established. In many cases, this may need to be integrated into existing change management activity within the data center. Key among these is tracking the changes made to all components of the storage network (see Chapter 23). It becomes particularly troublesome if you formalize changes to the switch configurations and not to upgrades of critical components such as HBAs, routers, and attached storage devices.
An important aspect of product installation is establishing a backout practice. Because the storage network is an infrastructure in and of itself, the reliability can be problematic in the beginning, as with any new technology. However, being able to back out quickly and return the production environment to an existing state saves valuable time as you move into a storage networking environment.