Information technology (IT) departments at state and local government agencies often are torn between competitive priorities - conserving budget dollars while making IT investments to improve agency operations. Fortunately, annual state and local IT spending will increase to approximately $33.5 billion and $28.9 billion respectively by 2009, according to market research firm Gartner Inc. - giving state and local IT managers the opportunity to use budget increases to invest in systems that cut costs and improve efficiencies. Blade servers reduce real estate and power costs while providing enormous flexibility and scalability over standalone servers, bringing real meaning to the adage "less is more."

Less Hardware Equals More Savings

Agency back-end expenditures are scrutinized closely. To win approval for budget requests, IT managers must try to improve operational efficiencies while maximizing savings to the agency. Blade servers, steadily growing in popularity, offer immediate and long-term savings.

Blade servers resemble a savings iceberg. On the surface, 20 percent of savings including density, power consumption and cable management can be found; just below are less visible savings from hardware, software and network platform integration, scalability, security and manageability. An organization using blade servers effectively can realize a 50 percent savings in operating and capital expenses.

In the near term, blade servers offer compelling capital cost savings to agencies undertaking routine technology refreshes. According to a study by Network World Lab Alliance, a typical blade system containing 14 blades, or servers, costs $117,400 to purchase. When accounting for power consumption, the total cost of implementing a blade server system is approximately $123,000. Alternatively, rack-mount servers with the equivalent computing ability and power consumption require more hardware and cost approximately $150,700 at the time of implementation.

In the long term, blade servers further reinforce the "less is more" adage. By packing as many as 16 blades (each 10U high; 1U represents 1.75 inches) into a standard 19-inch-wide chassis and using several chassis in racks of blades, it is possible to achieve a computing power that would typically require hundreds of stand-alone servers. What's more, each chassis of blades shares a single input/output (I/O) and power supply. By using less space and power, agencies can save on real estate and energy costs.

Because many blades are packed into a tight space, there is a common misconception that blades "run hotter" than comparable rack-mount servers. This is not the case, as blades generally have a lower wattage compared to rack mounts, which means they use fewer BTUs per hour. For example, the HP ProLiant BL460c blade server is a 350-watt server on average, while the comparable HP DL360 G5 rack-mount server is a 622-watt server.

Because blades are denser than rack-mount servers, best practices for air flow are required in the server room or data center in order to take advantage of lowered cooling costs. All blank spots in the rack should be covered with blanking panels, so the warm air emitted from the back of the blade enclosure does not circulate to the front of the rack. Hot aisles and cool aisles should also be established so that cool air continually moves through the enclosure and servers.

Less Hardware Equals More Utilization

State and local agencies with less-than-optimal server infrastructures often find it necessary to add servers faster than overall growth would warrant. The situation is similar to hosting a conference at a hotel. If all the rooms are the same size, some sessions are likely to overflow while others remain half-empty. Applied to server hardware, some servers are over utilized, while others are underutilized. To relieve overloaded servers, IT managers purchase additional servers, even though they may still have excess capacity overall. In this situation, IT managers grapple with time-consuming administrative and troubleshooting tasks as they seek to improve service levels.

This inefficient use of resources can be remedied by blade server architecture. With units made up of multiple server cards, blade servers can be adjusted for managing multiple applications, databases and storage devices, boosting utilization levels for servers previously used at only 25-30 percent capacity to 80-90 percent capacity. Because blade servers allow various applications and tasks to share central processing units (CPUs) and disk space, agencies can improve efficiency and power consumption with each application drawing its resources from a pool of servers. This also enables redundancy; with automatic failover, downtime can be eliminated and troubleshooting can be automated, which frees IT personnel for more high priority tasks.

Another efficiency factor to consider is that blade servers are easy to move. Without large amounts of real estate dedicated to servers, moving a physical technology system is much less of a hassle - whether it is part of a planned agency move or emergency equipment evacuation. Blade servers enable IT managers to move and reconfigure current equipment to a new location easily and without the need to purchase additional or new equipment, earning them excellent points for planning.

Less Really Is More

Moving to a new technology can be expensive and time consuming. This typically is not the case with blade servers, however. The newest generations of blade servers are relatively inexpensive and can be installed during routine technology upgrades. With downtime during server replacement as short as a few hours, the transition can be seamless, yielding immediate results. For an IT department in any state or local agency, blade servers answer the question of how to use less real estate and hardware to gain more efficiency and savings. ENS