Page 9 - TelecomReseller JunJul 2014

June/July 2014
Telecom Reseller 9
February/March 2014 20 Telecom Reseller
Data Center Downtime Study Uncovers Common
AUDIN
continued from page 1
Approaches to Reduce Unplanned Outages
RICH
obvious is the water consumption that is part of
the switch cooling process. An estimated 16.8
billion gallons of water will not be needed.
Using analytics to detect application
S
The electrical utilities emit CO2 emissions ONNER proven approaches to employing cooling, power performance anomalies
when generating the power for the legacy and management technologies to improve overall
Test generators and switchgear to
systems. The reduction in power consumption ensure emergency power.
By David Sonner, Vice President data center performance.
produces a 40% reduction in CO2 emissions.
The most rigorous form of this testing “pulling By Charley Rich, VP ProduMctaMrkaentiange,mLienbte,rNtaAsCtePlower,
Allocate sufficient resources for
problems arise, especially in today’s com- plex topologies the failure of a single appli-
Who Benefits?
the plug,” is mandated to meet local codes for Emerson Network Power (www.
IT organizations are under more pressure to
cation is rarely the culprit. Performance
emersonnetworkpower.com)
recovery in case of unplanned outages.
some industries. It confirms the proper operation Adeliver exceptional business performance threats usually are the result of multiple
The environment will benefit by the reduction in during a utility outage of the automatic transition than ever. Further complicating the chal-
Every data center should have detailed plans in
issues—and many of these, if caught early
water consumption and lower CO2 emissions. These two benefits though may end up as part
from utility to battery to generator and back. It keeps the facility team updated in their training, 2013 survey of U.S.-based data center
the event of a natural disaster. Recent disasters
lenge is the evolving nature of Information
in the process using real-time analytics,
professionals from the Ponemon
taught us that having enough generator fuel
of the provider spin for the IP transition as a should an unplanned outage occur. For the Technology (IT). The rise of Big Data, mobile,Inclsotiutudt,easnhodwBsYthOatDa mhavjoeriatyddofed com- could prevent much larger failures from
on hand and an established supply chain for
justification but are not the real reasons for the IP facility management team, this also confirms
prelsepxointyd,enmtsahkaivnegexiptereivenecredmaonreuncphlaanlnlendgdiantga replenishment was critical to some facilities
occurring. Evading cascading failures is
transition.
fcoenr tIeTr otuotaagceqinuitrhee tphaestv2i4simbiolnityhst.hTheye fnaecetsdartoe maintaining availability. Those efforts include
essential. Ideally, IT Specialists should
the data center will ride through a utility outage
dsoebtercitnag;nroemspaolnidese.ntsexperiencedanaverage
supportfromseniormanagementoneffortsto
withoutincident,andgivesthemtimeina Thereductioninpowerconsumption,reduced
avoid being in the position of putting out
Today, an organization’s application
fires—they should be able to make sure the
controlled manner to harden any deficiencies.
real estate, and water usage will be economically of two complete data center outages during the
fire never starts. But, without the necessary
attractive to the carriers. The reduced costs will infrastructure typically includes Web com-
past two years. Partial outages, or those limited
The growing reliance on
visibility this is no simple task.
Review or monitor UPS batteries.
help pay for the IP transition.
ponents, messaging middleware and main-
to certain racks, occurred six times in the same
To properly manage today’s application
Having a dedicated battery monitoring system is What I wonder is will the reduced legacy frames. Application performance is impact-
endvairtoanmceentn, toergranoizpaetiornastmiounstsbe able to
timeframe.
a must. According to Emerson Network Power’s OPEX cost translate to savings in the service cost ed by many factors coming from multiple
analyze the entire application chain from
The results do show that companies are more
for business profitability
Liebert Services business, batteries are the to the consumers and enterprises now on the sources—application servers, messaging
end to end, understanding the dependen- aware of downtime causes and are taking steps
protocols, virtualized systems, capacity
leading cause of UPS failure. Using a predictive PSTN? Will the carriers try to keep their service cies between the links in the chain. It must
has also impacted the site
to minimize the risk. The survey examined
issues and many more. Inevitably, failures
battery monitoring method can provide early costs high justifying the high rates to pay for the also be able to focus on early detection
high-performing data centers that experienced
selection process for new
in one or more of these systems occur—and
notification of potential battery failure. The best transition? Probably.
of abnormalities, differentiating symptom the least amount of downtime and identified five
IT is left to deal with the result.
data centers
practice is to implement a monitoring system I think the regulators should factor in the from cause rather than simply reacting to
common issues and actions to avoid outages.
Such situations are why Application
an outage. The combination of these two
that connects to and tracks the health of each significant OPEX savings to the carriers as Not every data center will be able to adopt all
Performance Management (APM) solu-
factors provides the level of assurance IT
battery within a string.
described above as an offset for the CAPEX of these traits; but even implementing a few could
tions exist. To be effective, APM must
needs in its key mission: to reduce the fre-
Implement data center infrastructure incurred by the providers for the IP transition. decrease the frequency and severity of unplanned
prevent and manage unplanned outages. The
deliver three major benefits: it must gain
quency and duration of outages.
management (DCIM).
I expect all this OPEX reduction will go to the downtime.
survey shows that respondents attributed an
enough visibility to see an entire system; track activities through the infrastructure End-to-end performance monitoring and
It is important to ensure the foundation for bottom line of the carriers not the bill to the consumers and enterprises.
cDhatina acsenthteyr aovcaciulra;bainlidtycoisrraelpatreioervietny.ts— average of 30 percent of unplanned outages to
weather related events. The growing reliance on
analysis must embrace the entire IT envi-
effective management of the data center is in Walsh has posted a blog, “Network mAsa7n1ypoerfcewnthoicf hPomneimghotn sresepmonudennrteslainteddic—ateidn data center operations for business profitability
ronment, from .NET to mainframes. It must
place in the form of an up-to-date visual model Transformation Is Everybody’s Business: What othredier rbutsoinesspmotoddeel ivsedleopeindgentroendthse before
has also impacted the site selection process
cover a wide range of components from
of the facility and centralized monitoring of Will Your Legacy Be?” - See more at: http:// users are impacted. Surprisingly, a number
J2EE application servers, Web Services to
data center to generate revenue and conduct
for new data centers - while utility costs were
middleware messaging, brokers and even
infrastructure systems. This will likely include www.genband.com/driving-innovation/network- of APM platforms miss on one or more of
commerce, dependability on maintaining
a primary consideration when planning the
legacy applications. It must also be elastic,
the deployment of a DCIM platform capable
transformation-everybody%E2%80%99s-business- these key functions.
uptime is more of a concern now than a year
location of a new data center, now flood plains
having the ability to transparently scale to
of providing a holistic view of data center what-will-your-legacy-be#sthash.ptozGFjc.dpuf To be sure, most APM solutions do a
ago. Availability can be maximized by adhering
and hurricane histories are carefully weighed
meet unexpected surges in demand.
operations based on real-time data that spans that summarizes the changes expected due to the good job of monitoring individual applica-
to best practices in data center design and
in order to ensure that Mother Nature will not
facilities and IT systems.
IP transition.
impair availability.
tions. But, monitoring is not enough. When
See RICH, page 21
redundancy. Established best practices represent

Use coupon code: DESIEC

Order online or by phone.
Expires: 10 / 1 /2014

p1-28_TR_June/July_2014.indd 9
11/07/2014 19:51