There's an old set of tradeoffs between buying services or "doing it yourself," where it comes to computing or communications infrastructure. Hosted VoIP virtually always makes more sense than buying systems for a smaller business. But premises-based solutions typically are more economical for large enterprises.
Something of the same argument can be when companies or people choose between cloud computing services and building their own data centers. Obviously, large enterprises often justify building their own data centers. Others might be able to justify renting space in somebody else's data center. Smaller organizations might well find that renting computing cycles is the better choice.
Google Sr. Manager, Production Network Engineering and Architecture at Google argues that the decision is highly dependent on duty cycle. Steady, predictable loads, especially at a high rate of utilization, will tip economics in favor of self-operated or co-located facilities. Highly-variable demand, and low volume, will tend to tip the economics in favor of a cloud computing solution.
"Think of it as taking a taxi vs. buying a car to make a trip between San Francisco and Palo Alto," says Gill. "If you only make the trip once a quarter, it is cheaper to take a taxi." But "if you make the trip every day, then you are better off buying a car."
"The difference is the duty cycle. If you are running infrastructure with a duty cycle of 100 percent, it may make sense to run in-house," says Gill. The detailed assumptions and analysis are here: https://spreadsheets.google.com/ccc?key=0AgWfa8v6EGzjdElXQVFzU1plSXdEQmVHZ3M5YjlsNVE&hl=en&authkey=CM_RzL0E#gid=0
link to post
Saturday, August 21, 2010
Cloud Computing: When to Use it; When Not To
Labels:
Amazon,
cloud computing,
Google,
Vijay Gill
Gary Kim has been a digital infra analyst and journalist for more than 30 years, covering the business impact of technology, pre- and post-internet. He sees a similar evolution coming with AI. General-purpose technologies do not come along very often, but when they do, they change life, economies and industries.
Subscribe to:
Post Comments (Atom)
It Will be Hard to Measure AI Impact on Knowledge Worker "Productivity"
There are over 100 million knowledge workers in the United States, and more than 1.25 billion knowledge workers globally, according to one A...
-
We have all repeatedly seen comparisons of equity value of hyperscale app providers compared to the value of connectivity providers, which s...
-
It really is surprising how often a Pareto distribution--the “80/20 rule--appears in business life, or in life, generally. Basically, the...
-
One recurring issue with forecasts of multi-access edge computing is that it is easier to make predictions about cost than revenue and infra...
No comments:
Post a Comment