Published: 17 Nov 2002
The key to an effective storage infrastructure is a dedicated group that is further subdivided to develop services, engineer them and operate them
Hearken back to IT about 15 years ago. Computing was restricted to the data center or a few lucky departments. Storage was thought of as a system peripheral, connected to a host system through proprietary channel technologies such as ESCON, IPI-3 or SDI. Because storage was thought of as a part of the system, storage management was placed in the domain of system administrators who were responsible for procuring, implementing, managing and protecting the data associated with each system.
Now fast forward. Today the evolution and proliferation of computing and networking technologies has led to new applications, distributed systems and users, and mountains of new data. To cope with these changes, IT organizations have evolved, creating network, database and application groups in the process. Yet, in spite of technology change and massive new storage requirements, storage management and administration is still assigned to system administrators in most IT shops.
|The storage services organization|
This organizational model has outgrown its usefulness and is no longer the best way to manage storage requirements. It's time for a drastic change.
The time to act is now
Storage requirements continue to grow 50% to 100% per year. And information is a critical corporate asset that needs to be managed and protected. Someone within IT must have the necessary skills to manage and protect it.
From an IT perspective, the transition from direct-attached storage (DAS) to network-based storage (SAN/NAS) presents CIOs with an opportunity to fine-tune the IT organization while meeting business requirements. And it presents storage specialists with a similar chance to further their own profession, since networked storage presents some new challenges to IT groups:
Network storage is a gray area. Storage area networks (SANs) and network-attached storage (NAS) don't really sit in the system administrator domain. They fall somewhere between the systems, networking and security groups, but also require particular skills around Fibre Channel, HBAs, volume management, and storage devices. This knowledge is too specific-and important-to be left as an afterthought to other IT groups.
Storage networks are designed for sharing. Fibre Channel switches, volume managers and LUNs let companies share storage systems to increase utilization, centralize backup/restore operations and get more out of staff. To do this effectively, storage architectures need to be designed, implemented and operated properly.
Data movement is complex and essential. Managing corporate data is no longer as easy as a database query or TAR command in Unix. Effective data management requires an array of skills in replication, backup, archiving and remote mirroring over both LANs and WANs. Poor processes in these areas could lead to costly downtime, regulatory problems or worse yet-lost data. This is no place to gamble-CIOs should make data management part of the storage services team chores.
Storage services groups can help companies get more out of equipment and staff to keep IT costs in control. And by offering storage as a service and improving data management, IT managers can provide better service to the business.
The storage services organization
Once a decision is made to form a storage services group, it should be focused on three areas:
Storage services management. This subgroup should create a menu of storage service offerings that meet the needs of internal customers and satisfy business requirements. Storage consumers should be given a choice of services based on their availability, scale, data protection and archival requirements. Vendor choices should be avoided-the important criteria should be limited to service level requirements.
Storage services customers pay for services based upon requirements-the more they need, the more they pay. The 7x24 e-commerce group that requires redundant HBAs, RAID devices, hot standbys and remote mirroring will pay a lot more for storage services than the NT file server team.
The storage service management group should track SLA performance and report back to customers to answer the obvious question, "What are we paying for?" And the group should drive continuous improvement.
Storage engineering. While the storage services management team is gathering requirements and creating services to meet them, the storage engineering team is actually piecing together the technology widgets to meet these needs. This group is responsible for storage research and development, design, test and implementation for all storage hardware, software and networking solutions. Interoperability is a key to success here as well-the team must constantly ensure that software revisions, firmware and patches are up to date. Storage engineering is responsible for coordinating with other IT teams when necessary. For example, when the data
center manager needs to mirror transactions across the Hudson River, it is up to storage engineering to work with networking to provision an OC-48 SONET circuit as part of the solution. Storage engineering is also responsible for utilization and capacity planning to keep ahead of demand and ensure that future needs are met rapidly.
Storage operations. Just as storage engineering designs and implements the storage architecture, storage operations ensure that everything runs smoothly. This process includes FCAPS (fault, configuration, asset, performance and security) management, along with operational processes, procedures and documentation. To improve data protection and operational focus, storage operations should be subdivided into two groups, one for online and one for offline storage. Online folks handle day today configuring and tweaking of the storage network and its piece parts while the offline team concentrates on backup/restore, remote mirroring, tape management, archival and vaulting.
By separating management, engineering, and operations functions, the storage services group will develop the expertise needed to provide excellent service and data protection for its customers. To evolve this process over time, you need ample communications between storage services groups and internal customers to address business needs, improve processes, increase efficiencies and lower costs.
As with any IT activity, it is important to measure the storage services group to make sure that it is continually improving and delivering great service. The success metrics can be divided into three buckets:
Problem resolution metrics. This group includes things like Mean Time Between Failure (MTBF), number of storage related events and trouble tickets, storage downtime, project delays, etc. An effective storage services group should proactively address issues and rapidly fix problems that do arise.
Cost metrics. These metrics follow equipment and personnel statistics like disk utilization, average cost of a megabyte and terabytes managed per storage services employee. Use what you have more effectively, pay less for new stuff and increase the amount of work each person does (without adding more work hours).
User satisfaction metrics. The ultimate goal of a storage services group is to provide better service to internal customers. Did the storage group provide a good package and timely delivery to applications development? Were file restores done quickly for marketing? Was data moved in a timely fashion to populate the new business intelligence system? If the answer to questions like these was a definite yes, the storage services team is doing its job.
To measure progress, perform a baseline assessment in all three areas before implementing a storage services team. This information will provide a starting place and also define specific problem areas. Once the baseline is established, measure progress on a monthly basis. Strive for continuous improvement. Report your findings within IT, to the business units, and up to senior management. Advertise your success.
The time to debate the merits of a storage group is over. Smart IT groups will establish a storage services group ASAP to improve service to internal customers, automate processes and cut costs. IT wins by improving performance and cutting costs while end users and business units win through improved storage services and better data protection.