When I speak with IT managers, I constantly hear "Yeah, we're moving to an IT services model as soon as possible."...
But when I ask what this means, the details become sketchy. IT is great for building technical solutions, so it will also be great at building technical services, right?
Wrong. There's a huge difference between a technical solution and a technical service. Technical solutions go through two phases. During the first phase, business and IT managers define requirements and develop a solution to meet these needs. Then IT takes over with design, implementation and ongoing management.
In technical services, the relationship is different from the start. IT sells a bundled service (in this case, a storage service) to the business units, complete with an impending commitment to service levels, meeting future needs and continually reducing cost. A technical service is a cooperative effort between a customer and a service provider.
To make storage services work, the storage services group within IT must complement its technical know-how with a service provider mentality. Service providers depend upon a set of common rules of engagement so they can provide predictable, cookie-cutter services. IT must follow some of these well-known practices.
Adhere to standards
When compared to a global telecommunications network, even the biggest IT infrastructure looks simplistic. How can carriers deliver telecommunications services to their customers while managing network complexity? Carriers standardize networking gear to maximize manageability, improve maintenance and increase vendor discounts. Staff become more knowledgeable and efficient if they can focus on a single system.
Storage services must follow this model. The storage services group should create standard gold, silver and bronze packages, depending upon performance, availability and data protection requirements (see "Rolling out storage services"). These packages should be composed of standard equipment, configurations and architectures. Under some circumstances, you may also want to offer a la carte pricing - customized packages of services to fit special needs. However, you should retain control of what goes on the a la carte menu.
The storage services team should base equipment selection on budget, vendor qualification and testing. Storage engineers must make sure that all equipment is tested for compatibility. Insist on testing a configuration that emulates each service package. Once standards are approved, exceptions must be discouraged, as each deviation means additional support and maintenance costs.
The storage industry is flirting with some technical standards such as the common information model (CIM) and Bluefin. Once implemented, these standards will help IT cut costs and improve operations. Make sure the vendors you choose have a credible plan for standards compliance as storage standards are blessed by Storage Networking Industry Association (SNIA) and other industry groups. Vendors should be contractually obligated to comply with standards in a defined timeframe or face stiff penalties.
Emphasize operations, not just technology
Service providers are big believers in Murphy's Law. In order to deal with unexpected problems and maintain service levels, operational processes and procedures need a lot of attention. This means that designing and developing storage services should be a cooperative effort between storage engineering and the storage operations groups.
The first place to focus this collaboration is on day-to-day monitoring and management. How will the operations team determine whether storage equipment is functioning correctly? This is a larger problem than it seems, as equipment management is still extremely proprietary and storage operations management tools are immature.
Given the limitations of current storage management software, your best bet is to use existing processes and technologies wherever possible. To capture events as they occur, integrate simple network management protocol (SNMP)-based storage devices into network operations center (NOC) event management systems such as Hewlett-Packard's OpenView or Micromuse's Netcool.
Typical NOC staff has limited experience with storage devices, so you'll need to provide them with the storage-centric training and tools so they can do their jobs. Document error codes so operators can quickly diagnose problems and tell the difference between a hiccup and a crash.
Make sure technicians can quickly identify whether an event impacts gold, silver or bronze storage services, and define escalation processes based upon the impacted type of service. If performance degrades in a bronze service, you may want to have a junior storage technician diagnose and solve the problem.
When the same condition appears in gold services, you'll want to escalate the problem immediately to the most qualified members of the storage services staff and have a detailed plan for problem isolation and repair.
You may also want to develop a quick and dirty storage Intranet to provide an elementary single pane of glass solution. This can be as simple as an HTML front-end or as sophisticated as an XML doc that integrates data from various sources. Remember to add a security layer to restrict access to this critical information.
Other important IT procedures to consider are change management, configuration management, including revision levels and patches, vendor management and security management. You'll need documented, stringent processes for all of these disciplines.
The customer always comes first
Service businesses such as restaurants live and die by this rule. If a customer is unhappy with his steak, a good waiter won't dwell on the qualifications of the chef, freshness of the meat or ingredients in the sauce, he will focus immediately on making the customer happy. The storage services group should take this restaurant example to heart. IT departments have often focused on technology rather than users. Since business units will actually be paying for storage services, the storage services team must reverse this behavior and concentrate on keeping its customers happy.
The first element to storage service customer bliss is communications. Business managers must understand what they get, what they don't get and how they'll know the difference.
Communication processes should also address outages. Users need immediate notification when services are impacted and when they will be restored. Announce planed downtime as well, even if occurs on a weekend. If you don't, you may have the e-mail storage offline just when the CEO is logging on to the VPN and attempting to fetch their e-mail messages.
Keep customers happy with accurate and appropriate reporting. Business users need simple answers to basic questions such as: Will I run out of storage this month? Is my data protected? No fancy real-time monitoring tools are needed here. Use the corporate Intranet to post basic graphs and charts on a daily basis.
Finally, there's billing. In general, business managers find IT charge-back bills to be one of life's great mysteries. Since they don't understand what they're paying for, they can never be satisfied. Smart storage service managers will seek out billing requirements and preferences at the onset of the project, before they commit any resources. When bills are mailed, storage services personnel should be ready to answer customer questions and make necessary modifications to preserve customer satisfaction.
Business people just want to pay for the storage they need and get what they pay for. If the storage services team empathizes with users, embraces a service philosophy and sticks to the rules I listed earlier, they are well on their way to improving service and cutting costs.