Data Centers: Part of the Cost Puzzle
August 11, 2008
The “commentary” is “Servers: Why thrifty Isn’t Nifty” which appears here. The “commentary” is by a wizard, Kenneth G. Brill, and he takes a strong stand on the topic of data center costs. The “commentary” is sponsored by SAP, an outfit that exercises servers to the max. Mr. Brill is the executive director of the highly regarded Uptime Institute in Santa Fe, New Mexico. Santa Fe is a high-tech haven. The Santa Fe Institute and numerous think tanks populate this city, a reasonable drive from LANL (Los Alamos National Laboratory). LANL is world famous for its security as you may know. With chaos theory and technical Jedis in every nook and cranny of the city except the art galleries, I am most respectful of ideas from that fair city’s intelligentsia.
The hook for the “commentary” is a report called Revolutionizing Data Center Efficiency. The guts of the report are recommendations to chief information officers about data centers. With the shift to cloud computing, data centers are hotter than a Project Runway winner’s little black dress. For me the most interesting part of this “commentary” was this statement:
One of these recommendations is to dramatically improve cost knowledge within IT…The facility investment required to merely plug-in the blades was an unplanned $54 million. An additional unplanned $30 million was required to run the blades over three years. So what appeared to be a $22 million decision was really an enterprise decision of over $106 million.
The “commentary” includes a table with data that backs up his analysis. The data are useful but as you will learn at the foot of this essay, offer only a partial glimpse of a more significant cost issue. You may want to read my modest essay about cost here.
What baffles me is the headline “Servers: Why Thrifty Isn’t Nifty”. Forbes’s editors are more in the know about language that I. I’m not sure about the use of the word thrifty because the “commentary” uses servers as an example of the cost analysis problem facing organizations when folks make assumptions without experience, adequate accounting methods, and a rat pack of 25 year old MBAs calculating costs.
Let me make this simple: cost estimations usually have little connection to the actual expenditures required to make a data center work. This applies to the data centers themselves, applications, or the add ons that organizations layer on their information technology infrastructure.
Poor cost analysis can sink the ship.
Mr. Brill has done a fine job of pointing out one cost hockey stick curve. There are others. Until folks like the sponsors of Mr. Brill’s “commentary” spell out what’s needed to run bloated and inefficient enterprise applications, cost overruns will remain standard operating procedure in organizations.
Before I close this encomium to Santa Fe thinking, may I point out:
- Engineering data centers is not trivial
- Traditional methods don’t work particularly well nor economically in the world of multi core servers and peta-scale storage devices stuffed into poor engineered facilities
- Buying high end equipment increases costs because when one of those exotic gizmos dies, it is often tough to get a replacement or a fix quickly. The better approach is to view hardware like disposable napkins?
Which is better?
[a] Dirt cheap hardware that delivers 4X to 15X the performance of exotic name brand servers or [b] really expensive hardware that both fails and runs slowly at an extremely high price? If you picked the disposable napkin approach, you are on the right track. Better engineering can do more than reduce the need for expensive, high end data center gear. By moving routine tasks to the operating system, other savings can be found. Re engineering cooling mechanisms can extend drive and power supply life and reduce power demands. There are other engineering options to exercise. Throwing money at a problem works if the money is “smart”. Stupid money just creates more overruns.
Mr. Brill’s “commentary” provides one view of data center costs, but I trust that he has the brand name versus generic costing in the report he references. If not, there’s always an opportunity in Santa Fe for opening an art gallery or joining LANL’s security team.
Stephen Arnold, August 11, 2008
Stephen Arnold, August 11, 2008
Microsoft SharePoint Olympic Watch
August 11, 2008
Microsoft’s plan to get Silverlight on millions of personal computers is now underway. It’s too soon to determine if it wins the gold for software downloads. One of my sources reports that the Mojave Web site runs on Flash. Hmmm. If this image is real and not Photoshopped, , I guess most attendees know what this translucent blue screen shot is. A BSOD (blue screen of death) appears at the Chinese Olympics. You can see the ghostly image here, courtesy of PowerApple.com. In case the image 404s, here’s what I saw.
If you have any additional information about this “image”, please, let me know.
Stephen Arnold, August 11, 2008
Hot News: Google Is Interested in Content
August 11, 2008
That wild and wonderful New York Times has a rear view mirror article that you must read. It’s here and called “Is Google a Media Company?” by Miguel Helft, a really good writer. For me, the key point in the article is this statement:
Google has long insisted that it has no plans to own or create content, and that it is a friend, not a foe, of media companies. The Google search engine sends huge numbers of users to the digital doorsteps of thousands of media companies, many of which also rely on Google to place ads on their sites.
This is, of course, Google’s standard verbiage, its “game plan” talk.
Mr. Helft quotes a range of experts who offer a contrary view. A Harvard professor (David B. Yoffie) surely is in the know, is quoted saying:
‘If I am a content provider and I depend upon Google as a mechanism to drive traffic to me, should I fear that they may compete with me in the future?’ Professor Yoffie asked. ‘The answer is absolutely, positively yes.’
I talk a bit–I recall I devoted 20 or 25 pages–to Google’s publishing and content acquisition / distribution inventions in my August 2007 study Google Version 2.0. If you are curious, there’s more information here. Outsell, a nifty consulting outfit in Burlingame, California, recycled some of my research late last year. There is a bit of dissonance between what my research suggested and the tasty sound bites in the New York Times article.
The key point is that Google’s been beavering away in “publishing” for quite a while. Actually, publishing, even the word media, is too narrow. Google has somewhat wider vistas in mind if I understand its patent documents and technical papers.
It’s exciting to know that now the paper of record has made it official. Google has some media thoughts in its Googzilla brain.
Stephen Arnold, August 11, 2008
Google and Hosted Telephony
August 11, 2008
Network World’s Matthew Nickasch wrote an interesting article “Will Google Consider Hosted Telephony?”. You will want to read it in its entirety. The story is here. The premise of the story is that Google may offer a range of wireless services from the cloud. Mr. Nickasch asserts:
While no official plans, or even rumors have been released, a Google-hosted VoIP environment may be incredibly popular for organizations that utilize Google Apps for all other collaboration needs. We’ve seen our fair share of free hosted VoIP environments, like Skype, Free World Dialup, etc, but Google has yet to venture into such a market.
My own research into Google’s telephony activities suggested to me that:
- Google started working on mobile and other telephony services as early as 1999
- Telephony, based on my analysis of Google patent documents, has been one of the areas of intense activity for almost a decade.
- Google’s innovations extend deeper than hosted applications; for example, Google has a clever invention for routing calls in a distributed mesh environment.
Mr. Nickasch ends his article with several questions. What’s your take? Has Google lost its chance to make a telco or has Google a different game underway? In Google Version 2.0, I discuss options for Google’s “other game”. Hosted services are already here, and I think Googzilla is watching and learning.
Stephen Arnold, August 11, 2008
Beyond Search’s Search Function Back On Track
August 10, 2008
I have had many positive comments about the search function for my Web log “Beyond Search”. Last week, we had reports of current postings not appearing in the index. Our hosting company had in place a method to block certain clickstreams when certain conditions were detected by the hosting company’s automated systems. The increasing demand for access to the site and the additional content indexed by the Blossom search system caused a slow down in “Beyond Search.” The hosting company, Blossom.com, and my engineering team have resolved the problem. Thank for your patience. Blossom.com’s Web log indexing system continues to delight me. If you are looking for a search system for a Web site or a Web log, please navigate to http://www.blossom.com and check the company. Feel free to mention that Beyond Search is happy. I’m sufficiently happy to award the Blossom.com team three happy quacks. We’re back to normal, but my normal may be different from your normal. Anyway you can search for posts about SearchCloud, Sprylogics, and of course my favorite SharePoint. Enjoy.
Stephen Arnold, August 10, 2008
NTT Video Search
August 10, 2008
Video search evokes thoughts of Autonomy and Google, rarely NTT, the Japanese communications giant. According to DigitalBroadcasting.com, NTT has a robust media search technology cleverly named Robust Media Search. You can read about the invention here. The write up is a news release, so it has a pro-NTT spin. Imagine that.
NTT is no newcomer to search. The company has been pecking away at media search, based on proprietary NTT technologies, since 1996.
Video search is important. In my own research, I find that many organizations plop a 20-something in front of a Web cam and capture minutes of video about a topic. Google, the search wizards, are among the worst offenders. Google records presentations accompanied by almost unreadable visuals about many topics. To find the occasional gem, I have to let the video drone as I listen for an interesting fact. The video search engines are not too good. There are false drops and often the need to run the entire video to locate the single point referenced in the search system. Grrr. I hate video.
Will the NTT invention make my life easier as I try to cope with the rising tide of rich media? If you want to learn more about NTT’s technology, you can navigate here and deal with the registration process. I refused to do this.
Here’s what I have pieced together about this new search technology:
- Search is one component of a number of rich media services. These range from distribution to digital fingerprinting. More about this is here.
- The wizard identified with the search technology is Takayuki Kurozumi, Ph.D. Bio here.
- A field trial with BayTSP began in April 2008. More about the trial is here. Information about BayTSP is here. The “TSP” is an acronym for “track, secure, and protect”.
I have not been able to locate public information about the outcome of this test. Based on my experience with Japanese search systems, the technology may find its way into a network service. The “search” does not necessarily mean that I will be able to look for a video. The “search” may be a function for a copyright holder to locate and track video or audio content used without permission of the copyright holder.
Stephen Arnold, August 10, 2008
Ads Soften, Google Crumbles: Doomsday Approaches
August 10, 2008
Investor Business Daily’s Pete Barlas reports that online advertising may be crumbling. His story “Survey Indicates Economy Even Taking Toll On Search Ads” relies on data from a Covario survey that supports the assertion “search ad spending last quarter rose by the smallest percentage since at least the start of 2007.” You can read the Investor’s
Business Daily article here. In the best tradition of good news and bad news, Mr. Barlas reviews both sides of the Covario’s research findings. I’m a bad news type, and my mind considered the impact of sharply lower ad spending on Google. Because Google has the lion’s share of the online ad market, any downturn kicks Googzilla in the shin. A big enough downturn, Google could experience periostitis. The problem is not fatal, but it may slow the giant, giving Microsoft and other competitors an opportunity to make headway.
Stephen Arnold, August 10, 2008
Google: More Cause for Doubt
August 10, 2008
SFGate.com offers interesting views of business actions. The article “The AOL Flub Has analysts Revisiting Google” delivers on two counts. First, Ryan Kim summarizes Google’s admission that its investment in America Online has lost value, lots of value. Second, the write up rekindles the ashes of Google’s attempts at diversification have failed. You can read the August 9, 2008, story here. Mr. Kim revisits Google’s scattershot product development and reminds the reader that Google has been distracted by investments in companies such as YouTube.com, which has become a magnet for litigation and a challenge to monetize. Google may have overpaid for such properties as DoubleClick.com and gobbled small companies and done nothing to make them grow. More troublesome is Google’s interest in technologies unrelated to its core business; for example energy and space travel. For me the most important point in the article was this statement:
“Other than search, what has Google done right? They have 1,001 products in beta, but what’s been successful?” Chowdhry {\[an analyst quoted by Mr. Ryan] asked. “There has been a sequence of missteps and failures, and this is not the end. They miscalculated the valuation of AOL, and this is the first time they’re admitting to it.”
Google has a dominant position in Web search and advertising. The company has a track record of success in online advertising. Is it now time to reassess Google as company with a single business model and little else?
Stephen Arnold, August 10, 2008
Sinequa Inks OEM Deal with Oxaproc
August 9, 2008
Sinequa’s information retrieval solution will be integrated into Oxalys Technologies into Oxaproc, according to ITRNews.com. Oxaproc is an e-procurement system. Antoine Renard, director of development at Oxylys said that ease of integration was a factor in the company’s decision to license Sinequa CS. OEM deals are highly prized among search and content processing vendors. A typical deal involves up front cash and a royalty. Once an information retrieval engine has been embedded in an enterprise application, ripping and replacing can give engineers and customers migraines. You can learn more about Oxalys here. Specific information about Oxaproc e-procurement is here. Details about Sinequa are here. You can read the interview with Sinequa’s Jean Ferré top gun here.
Stephen Arnold, August 9, 2008
Intel: What Business Is It In?
August 9, 2008
Intel’s push in cloud computing strikes me as a “me too” response to a customer rebellion that is brewing. Maintaining servers, struggling with heat and power consumption costs, and the the mind-numbing wackiness of enterprise software fuel the shift. Intel in a search for more revenue is looking for a two-fer.
Intel wants to grow its revenue, particularly in its semiconductor business and Intel wants a bigger piece of the action in cloud computing. Can Intel perform this trick? This is a difficult question to answer. Now Intel seems to be probing other markets as well.
On August 8, 2008, Intel surprised me with its release of its Summary Statistics Library. You can read the Web log post by Dmitry Kabaev here. You can download the library here. There is also an installation guide available from the download page. You can choose either the Linux or the Windows library. There are two low key requests for your email, but as far as I could tell, I was able to suck down the libraries without registering. If you want to participate in Intel forums, you will have to cough up some information, but I register my dog, who seems quite happy to ignore his email.
The stats pack is part of Intel’s Whatif.Intel.com initiative. Intel wants to be a good open source citizen, and it is an excellent way to allow developers to start mud wrestling with programming for massively parallel systems. Intel is upfront about this point, describing the library as “a set of algorithms for parallel processing of multi-dimensional datasets. It contains functions for initial analysis of raw data which allow investigating structure of datasets and get their basic characteristics, estimates, and internal dependencies.”
You can whack on data sets with:
- Basic statistics. Algebraic and central moments up to 4th order, skewness, kurtosis, variation coefficient, quantiles and order statistics.
- Estimation of Dependencies. Variance-covariance/correlation matrix, partial variance-covariance/correlation matrix, pooled/group variance-covariance/correlation matrix.
- Data with Outliers. The Intel® Summary Statistics Library contains a tool for detection of outliers in a dataset. Also the library allows computing robust estimates of the covariance matrix and mean in presence of outliers.
- Missing Values. Data which contains missing values can be effectively processed using modern algorithms implemented in the package.
- Out-of-Memory Datasets. Many algorithms of the library support data which cannot fit into the physical memory processing huge data arrays in portions. Specifically, variance-covariance matrix estimators, algebraic and central moments, skewness, kurtosis, and variation coefficient can process a dataset in portions.
- Various Data Storage Formats. The Intel Summary Statistics Library supports in-rows and in-columns storage formats for datasets, full and packed format for variance-covariance matrix.
The libraries support C and Fortran90/95.
Intel has invested in Endeca, and I don’t think this is a casual greenfield seeding. Endeca’s technology performs some interesting processes on structured and unstructured content. I see not overt evidence that Intel is overtly moving into information retrieval. I am tracking announcements like this stats pack as part of my research effort to figure out how Endeca figures in Intel’s plans.
While I root around for information, download the statistics libraries. My quick look revealed some useful work by Intel’s engineers, who merit a happy quack.
Stephen Arnold, August 9, 2008
-
- Subscribe to Beyond Search
Feature archive
News archive
-
