If you like simplicity when it comes to considering which data center availability “standard” you want to follow, the 1-2-3-4 rating system may be just the right approach for you. However, these fixed steps restrict ways to consider variations in availability schemes, such as increasing the redundancy of the power chain to 2(N+1), while keeping the cooling system at N+1. Therefore, the 4-level system does not acknowledge that as the proverbial (but forbidden) 3+ rating. Unfortunately, this provides little or no flexibility for determining how well this reflects the interrelationship of your computing architecture with the facility infrastructure requirements.
Then consider taking a more holistic approach — The Green Grid Open has recently released the Open Standard for Data Center Availability (OSDA) in the form of an online tool to allow data center owners, operators, and users to project the overall availability of a data center. The OSDA project began in 2016, however this year I had the opportunity to be involved with part of the online developments. While I did not do the “heavy lifting” of applying the Monte Carlo Simulations, I can share that the workgroup put a lot of thought into, and many, many hours discussing and refining, the OSDA tool.
While the tool is still a draft version, it is functional and the Monte Carlo Simulations are used to interactively “model” the interrelationship and availability impact of “where” in the power chain you may want to beef-up or reduce your redundancy. For example, it allows you to consider the “reliability” factor of the utility power; either based on IEEE standards or using custom inputs based on historic conditions at the local level, to model failure domains. It also can provide insight in the effect of using an energy storage system to mitigate the impact of short-term utility outages.
The bulk of the OSDA development work done by the workgroup team was focused on the power chain. However, the cooling section is still in the early stages of development. Nonetheless, the basic cooling redundancy choices (N, N+1, 2N etc.) are included and functional.
To be clear, the tool itself is not meant to be a “certification” mechanism, but it will provide a different view of the weight of intertwining elements which could impact the “availability” of a design in a non-linear fashion. And while the OSDA scale pays heed to the basic four-step precepts and terminology (starting with Basic Non-Redundant and going toward Concurrently Maintainable/Fault-Tolerant), the results are displayed on a scale of 0-10 (compared to 1-2-3-4).
The OSDA evaluation methodology offers an inclusive, non-proprietary, flexible tool providing a more granular rating scale. This provides clearer high level availability insights than the rigid, 4-level classification systems traditionally used to rate a data center facility, which are based primarily on the redundancy of power and cooling infrastructure.
The OSDA initiative brings a less prescriptive framework to the data center community, allowing easier consideration of alternative energy sources and innovative efficient designs. The easy to use OSDA tool will help weigh the trade-offs of these type(s) of site design choices. Moreover, the OSDA tool also incorporates the ability to going beyond the four walls of the facility and can project the improvement of application availability by considering multi-site computing architectures and data replication as an option.
Clearly, those in the traditional data center universe are risk adverse and not very open to betting on new standards, and generally slow to consider adopting them. Along those lines, the majority of the industry closely follows the ASHRAE Thermal Guidelines, yet there were somewhat mild reactions to the release of the 2nd edition in 2008 that raised the “recommended” temperature from 77°F (25°C) to 80.6°F (27°C), and most operators did not change anything.
Moreover, the 3rd edition, which was released in 2011, produced a somewhat shocked reaction by introducing the expanded “Allowable” environmental ranges: A1 through A4, as well as directly using airside for free cooling. The mere mention of directly using airside for free cooling really rattled a lot of ASHRAE followers who would never think of letting in outside air (other than minimum fresh air required by code). The expanded allowable IT temperature ranges also caused doubt in those that firmly held the long standing rigid rule of “colder-is-better” and strictly enforced “don’t move the thermostat past 68°F.” As a result, despite the ASHRAE’s updated “Thermal Guidelines,” very few enterprise or colocation operators raised their operating temperatures much, if at all, for the first few years after it was published.
Of course, there are still many data center operators today that are still mentally uncomfortable with the concept of operating at or near 80°F in the cold aisle and still like “to keep their cool” (perhaps moving it up to 72°F instead of 68°F). Furthermore, many people just seemed to have just stopped reading and closed the book when they saw the upper bound of the allowable A4 IT equipment intake range at 113°F (45°C), not to mention 8% to 80% RH range listed in the 2015 4th edition.
Are you Feeling Lucky? Try the X-Factor
I teach several classes a year on a variety of aspects of data centers including optimizing cooling and energy efficiency to a broad range of highly experienced people. While they are all familiar with the “recommended” range of the ASHRAE guidelines, fewer give much consideration of operating in the “allowable” ranges. However, when I present and discuss the “X-Factor,” they react as if they have entered an alternate reality field. In case you may not be aware of the X-Factor, ASHRAE introduced it in the 3rd edition of the Guidelines and defined it as:
“x-factor, time-weighted (or net): a dimensionless metric indicating a statistical equipment failure rate over a defined range of environmental temperatures when compared to a constant baseline temperature of 20°C (68°F)”
This new “dimensionless metric” openly encouraged the use of non-mechanical (compressor-less) “free cooling” using direct outside air economizers to take maximum advantage of ambient air temperatures (based on the newly expanded allowable limits) to cool data centers.
This seeming radical declaration by ASHRAE, the risk adverse industry stalwart, was the result of the information the IT equipment manufacturers who internally shared their projected failure rates vs. temperature over the expanded temperate ranges. As a result of this anatomized internal data, they created the X-Factor risk projections. The publication of the X-Factor was meant to allow data center managers to make their own decisions as to how much risk they would like to consider to save energy, while still maintaining “acceptable” risk expectations, based on projected IT equipment failures.
The Bottom Line
To put things in perspective, it took a while for data center operators to even consider raising the temperatures to save energy, and now many new sites are designed to utilize free cooling and safely operate their IT equipment at 72°F- to 76°F. Although released in 2011, the X-Factor is still highly debated (at least among those who read it) and sometimes misinterpreted, and it continues on in the 4th edition of the ASHRAE Thermal Guidelines.
Like the X-Factor, the ODSA Tool provides more flexible examination and granular calculations of reliability risks. Data centers, IT equipment, and computing architecture have changed over the more recent years and will continue to evolve. The related metrics must adapt to those changes to remain meaningful. So just as ASHRAE updates the Thermal Guidelines to address these changes, consider examining OSDA tool’s unique approach to analyzing “availability.”
Now for the best part, while you will have to purchase the ASHRAE Thermal Guidelines to understand the full details of how the X-Factor risks are calculated, access to the OSDA on-line tool is now free, courtesy of The Green Grid.
To learn more on how to help examine the “odds” for improving your data center's availability, access the free OSDA tool at https://bit.ly/2QLBP88.
