FOCUS CLOUD UTILIZATION
Issue 10, June/July
PROVIDERS’ SUCCESS Economies of scale and treating all data center-related costs as a whole create savings in the cloud
centers just like other companies are operating data centers, how are they able to offer savings to their customers and still make a profit?
P
While part of the answer lies in the type of infrastructure they operate, the main cost reduction comes from the way they use their infrastructure and the way the cost of building and operating their infrastructure fi ts in their overall business model.
‘MISPLACED INCENTIVES’ IN TRADITIONAL IT PURCHASING
A study conducted in 2009 by the US Department of Energy’s Lawrence Berkeley National Laboratory, in collaboration with Microsoft, Intel and HP, analyzed the true cost of buying and operating servers. According to the paper outlining the study’s results, cloud-computing service providers have been ahead of the curve in their awareness of a relationship between indirect costs and the overall economics of IT. By “indirect costs” the researchers meant data center power and cooling expenses. “If power per server-cost continues to increase, the indirect power- related infrastructure costs will soon exceed the annualized direct cost of purchasing the IT equipment in the data center,” the report, titled Assessing Trends Over Time in Performance, Costs and Energy Use for Servers, reads.
The study warns that only taking in account performance per $1 of server acquisition cost will lead to overestimation of benefi ts investment in additional compute power will bring. As a general rule, performance per server and performance per $1,000 of server purchase cost double about every two years. However, performance per $1 of server cost has been increasing quicker than performance per 1 W of power in every case the study examined, which means power use per one unit of server cost has been increasing. The substantial capital cost of
28
www.datacenterdynamics.com
roviders of cloud-based services say that outsourcing IT functions to them will enable companies to reduce cost. They are operating data
power infrastructure also needs to be taken into account. At the time the report was published, the cost associated with power for cooling, back-up and distribution infrastructure came to about $25,000 per 1 kW of IT power use on average. Thus, while buying IT equipment with higher power draw lowers capital IT cost, it raises electricity cost and the capital cost of power-related infrastructure. (See Annual O&M costs graph on p.39).
Researchers acknowledged that cloud providers were ahead of the rest in their awareness and fi xing of these “misplaced incentives” by taking a more holistic look at their data center cost structure. Still, the paper says even these companies’ response to such awareness varies in sophistication. The misplaced incentives arise when IT and facilities departments have separate budgets or when colocation customers are charged on a per-square-foot basis, without taking into account their power usage. Possible fi xes for this are combining facilities and IT into a single cost model or outsourcing some compute capabilities to cloud service providers, who by the nature of their business model can deliver IT performance at a lower cost. Jonathan Koomey is a project scientist at LBNL and one of the lead researchers in the study. He says organizations should expect to realize bigger savings when outsourcing to public cloud providers, rather than building their own private clouds. “The cloud- computing folks, they have it all boiled down to one budget,” where the totality of OPEX and CAPEX of an organization is that of owning and operating their data centers.
Public cloud providers are companies like Amazon, Google, Microsoft, Rackspace or Savvis, who provide IT infrastructure as a service, where pricing is based on the amount of compute power clients use and the amount of time they use it for. To be sure, there is a multitude of various types of cloud-based services on the market, but this article focuses on compute capacity, or Infrastructure-as-a- Service, a Public Cloud based.
INCREASING INFRASTRUCTURE UTILIZATION RATES IS KEY TO CLOUD
THE MORE THE MERRIER
Besides greater awareness of the true cost of compute power, the two key reasons public- cloud providers can offer savings to enterprises and still make a profi t are economies of scale, heterogeneity of users of their infrastructure and focus on the business of delivering IT services, Koomey says. Cost per unit of computation is a lot smaller to a provider with a large-scale infrastructure than it is to a company with one or two data centers, whose core business is something other than provisioning of IT services. One of the key reasons is that any operational effi ciency improvements that are added to the model are applied at a wide scale and create large savings.
Improvements in things like infrastructure monitoring, power management, management of cooling capacity and better tracking of data center costs all have fi xed cost. “The more computation you can do, the more you can spread that fi xed cost over a larger number of computation (resources),” Koomey explained. Public cloud providers are able to increase the amount of computation their infrastructure performs – and increase utilization of capital invested in building and running that infrastructure – not only by having a large amount of customers but also by having a heterogeneous customer base. The wider the variety of customers, in terms of both type of business and geographical location, the higher the utilization rate will be.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84