|
 |
|
General Information
|
Add To My Personal Library |
May 5, 2006
• Vol.28 Issue 18
Page(s) 30 in print issue
|
How’s & Why’s Of Data Center Tiers
There’s Much More To Consider Than Just Raw Computing Power
|
When it comes to data centers, being No. 1 is not the best. Data centers typically are ranked in tiers, with Tier I the most basic and prone to downtime and Tier IV the most robust, redundant, and functional.
There are several infrastructure elements to consider beyond raw computing power and the number of fiber-optic pipes in and out of the data center. These site infrastructure features include power, cooling, and emergency backup capacity and functionality, height of the raised flooring, fire suppression, and both logical and physical security.
Tier I data centers have nonredundant power and cooling infrastructures. Tier IV data centers have all the bells and whistles—everything needed to keep them running without ever shutting down for maintenance, no matter what happens. The question for most small to midsized enterprises is where along the continuum they want their data center to fall. As always, the trade-off is between dollars and sense, although legal requirements such as Sarbanes-Oxley will weigh in, too.
 Tier Classification
The Uptime Institute (www.uptimeinstitute.com) created the Tier Classification system as a benchmark for reliable data center infrastructure design.
Specifying a tier level for an SME is like buying a suit or dress. Each outfit has different requirements and different costs. Some occasions require a tux or gown. For others, slacks will do. Setting up a one-size-fits-all standard will not work—either for business clothes or a data center.
Even sister data centers will have different requirements. For example, remote sites probably will have lower requirements than the main office, depending on how vital their role.
The tier levels give an idea of where along the continuum a data center might fall. But it is up to management to specify just how reliable a center has to be. Can you live with a few hours downtime every month? A few minutes? None?
Here Are The Breakouts:
Tier I: A single path for power and cooling distribution, without redundant components, providing 99.671% availability
Tier II: A single path for power and cooling distribution, with redundant components, providing 99.741% availability
Tier III: Multiple active power and cooling distribution paths but only one path active, redundant components, concurrently maintainable, providing 99.982% availability
Tier IV: Multiple active power and cooling distribution paths, redundant components, fault-tolerant, providing 99.995% availability
It is unlikely that the typical SME will need Tier IV redundancy unless it is involved in a highly specialized area such as banking or health care, where regulations such as the Sarbanes-Oxley rules apply.
Any SME can look at the table and determine about where their operation falls. Unless there is a legal reason or business case to certify to a given level, being familiar with tiers is a nicety, but not a necessity.
Tier Certification
If your operation is accountable to someone beyond the corporate suite, formal tier certification is probably a prudent investment.
The State of Oregon’s Information Resources Management Division got its Tier III certification in April 2005. Tier III standards certify the facility as a vast improvement over its earlier data center environment. From the beginning of the design process, it was determined that the state data center meet Tier III standards to provide concurrently maintainable, 99.982% availability.
Any one of the state’s new data center cooling units can be shut down with no impact to the computer equipment. Utility power can go down, and the data center will still have full internal power.
“Power lines and one of the generators can go down, and we will still have power to run the data center at full load,” says Dave Howard, service delivery manager for the state’s data center.
“One chiller can go down; the other picks it up. Take out one of the chilled water loops used to cool the facility, and the cooling system remains fully functional,” he says.
All Oregon agencies were mandated to provide facilities that were “minimally adequate” for the function required. The state data center was built as a new facility to meet Tier III standards. Howard says they brought together the architect, engineers, contractor, and consultants, and the result was a design/build project that provided what the state needed.
Where You Fall
An SME can get some idea of where its data center fits on the overall scheme of things by noting that Tier I-class data centers first appeared in the 1960s. About a decade later, standards were raised to current Tier II levels. Tier III came about in the late 1980s and early ’90s. The first Tier IV data center was developed in 1994 as part of the United Parcel Service’s Windward Project. It was the first site to assume the availability of dual-powered computer equipment and cost UPS $50 million to build.
A typical SME with between 350 and 700 seats likely has a data center that is Tier I or Tier II. A Tier I data center is susceptible to disruption from both planned and unplanned activity. This is the basic, default rating, says Julian Kudritzki, certification manager of the Uptime Institute.
Tier I sites have computer power distribution and cooling but may not have raised floors, UPSes, or engine generators. The critical load on these systems is up to 100% of N. Even with a UPS or generator, they likely are single-module systems and have many single points of failure. The infrastructure should be completely shut down on an annual basis to perform preventive maintenance and repair work. Urgent situations may require more frequent shutdowns.
Tier II centers have raised floors, UPSes, and engine generators. Their capacity design is N+1, which has a single-wired distribution path throughout. Maintenance of the critical power path and other parts of the site infrastructure still requires a processing shutdown.
In a Tier III center, most functions, including preventive and programmable maintenance, repair and replacement of components, and testing of systems, can be done without disrupting operation of hardware systems.
Tier IV is the best going, but even a fault-tolerant and concurrently maintainable Tier IV site does not meet the celebrated requirement of five nines (99.999%) uptime. The best a Tier IV site is expected to deliver over time is 99.995% reliability.
The Process
The time it takes for the certification process varies depending upon whether it is an existing or new site, Kudritzki says. For existing sites, two comprehensive assessments are involved. One is an onsite review of electrical and mechanical infrastructure (Continuous Availability Review), and the second is a detailed review of human factors (Site Infrastructure Operations Review). Following the site reviews, the Institute delivers detailed reports outlining what tier level can be achieved. Once the client has implemented the required modifications, a return visit is scheduled to confirm tier compliance and award site certification.
“One major variable is the responsiveness of the client’s team,” Kudritzki says. Some reviews done on existing facilities in 2004 are still awaiting the upgrades required to achieve the desired tier level.
A major factor is the money and the risk involved to upgrade an existing operating data center. According to Kudritzki, “It is much cheaper and easier to use the Institute’s Prospective Tier Classification process before starting construction of a new data center.” Moving valves, walls, and connection points or changing capacities and redundancies is less expensive and easier during the design process. The results in extended life and life cycle operational savings of prospective certification are profound.
A data center that wants to go through the process has to use a consultancy such as the Uptime Institute, unless the data center has someone on staff who has been through the process. As creators of the Tier Classification system, the Institute owns it. There are imitations, but if an SME wants a solid evaluation, they’ll either have to rent the expertise by using consultants such as those at the Uptime Institute or hire the expertise onto the payroll.
Budget for between $60,000 and $100,000 for a review from the Uptime Institute but keep in mind the figure ranges depending on the complexity of the project and uptime objectives.
“Going through the process depends on your needs,” Howard says. “Ours are high, with public safety and law enforcement on our program. You have to define your requirements. But if you need a top-quality data center, this is the way to go.” 
by Curt Harler
View the chart that accompanies this article.
(NOTE: These pages are PDF (Portable Document Format) files. You will need Adobe Acrobat to view these pages. Download Adobe Acrobat Reader)
View the chart that accompanies this article.
(NOTE: These pages are PDF (Portable Document Format) files. You will need Adobe Acrobat to view these pages. Download Adobe Acrobat Reader)
|
|
