ASHRAE Journal - November 2014 - 85

COLUMN DATA CENTERS

1.3

Class A2

1.2
1.1
Failure Rate at 68°F (20°C) - 24/7/365

Graphic reformatted by DLB Associates

1.0
0.9
0.8

.
t t l e Wi s o s t o n n v e r g e l e s i c a g o O h i o g t o n l a n t a a l l a s u s t o n o e n i x i a m i
D Ho
B
Sea ison,
M
At
De s An
Ch bus, ashin
Ph
ad
W
Lo
a
lum
M
S
o
C
FIGURE 2 Time-weighted x-factors for various U.S. cities. (Weather data from ASHRAE Weather Data Viewer.)
cis

©ASHRAE

ran
nF

LOCATION

% ANNUAL
HOURS AT
95°F

AVERAGE
X-FACTOR AT
95°F

% ANNUAL
HOURS AT
104°F

F = (B × C) + (D × E)

Location 1

1.55

1.66

0.0321

Location 2

1.55

1.66

0.0310

Peak Temperature Impact
A particularly high peak temperature is often cited
as being an inhibitor to a variable inlet temperature
approach, but since the peak temperature occurs for
a very short duration and for very few total hours in
a given year, its impact is minimal. Table 4 shows two
generic locations with different peak temperature
distributions and the net x-factor contribution difference between the two is roughly one-tenth of one
percent.

Some Words of Caution
The x-factor is predicated upon IT equipment inlet
conditions. If there are recirculation airflow problems
within a given cabinet or lineup, the actual inlet temperatures may be higher. Steps should be taken to properly
measure the inlet conditions on the servers and manage
the overall airflow patterns (e.g., consider containment
strategies such as blanking panels, etc.) to mitigate the
impact of recirculation.
As mentioned earlier, there are many types of IT
equipment and each of them may have a different failure rate profile, but the x-factors are based on the most

AVERAGE
X-FACTOR AT NET X-FACTOR CONTRIBUTION
104°F

common IT equipment in data centers: volume servers.
Routers, switches, and other types of servers may have
very different failure rate profiles from the data provided on volume servers. It is recommended that those
relative failure rates be categorized prior to aggressive
changes to inlet temperature conditions.
Last, the effect of a rapid rate of change of temperature
on failure rates has not been well studied or published at
this point. Changes in temperature to follow the Thermal
Guidelines' Recommended or Allowable envelopes should
be conducted gradually and in accordance with the latest rates of change clarification posted on TC 9.9's website (http://tinyurl.com/nsgb4cs).

Closing Comments
The release of relative failure rate data for IT equipment (x-factors) quantifies the impact of ambient inlet
temperatures on the reliability of IT equipment (for volume servers only). This information allows data center
owners and operators to make business-centric decisions on how to operate their data centers and establish
the right balance between reliability and energy-efficient operation using their own unique value system.
N OVEM BER 2014

ashrae.org

ASHRAE JOURNAL

Table reformatted by DLB Associates

TABLE 4 Generic location example of peak temperature impact. (Weather data from ASHRAE Weather
Data Viewer.)

©ASHRAE

at a given temperature with the associated
x-factor for that temperature. Table 3 shows
the time-weighted average x-factor calculation for Atlanta.
Even though Atlanta may be considered
one of the hotter climates in the U.S.,
its time-weighted annual failure rate is
1.052. When considered against the typical
annual server failure rate of 2% to 4%, this
represents an increase of 0.01% to 0.02%
compared to operating a precise cooling
inlet temperature of 68°F (20°C) (i.e., the
annual server failure rate becomes 2.01%
to 4.02% as opposed to 2% to 4%). For all
intents and purposes, this is a negligible
increase.
Figure 2 shows time-weighted average
x-factors for a number of U.S. cities. Many
cities have time-weighted x-factors below
the baseline of 1.00, and even one of the
warmest cities (Miami) only has a timeweighted x-factor of 1.25 (or an annual
server failure rate of 2.5% to 5%).

ASHRAE Journal - November 2014

Table of Contents for the Digital Edition of ASHRAE Journal - November 2014