Google’s Plumbing
April 8, 2009
Short honk: Don’t think the GOOG has a lead in data centers? Check out this video available from Google Blogoscoped. Keep in mind the factoid from my 2005 Google study here. Google’s $1.00 of data center investment returns more than $4.00 in performance compared to high end, branded servers from big name manufacturers. Cheap, fast, and scalable. Google’s plumbing delivers throughput at a lower cost than the competitions’ data centers. What’s this mean in capex? A competitor’s data center costing $650 million would have to be four times larger to match a comparable Google data center’s performance. That’s more than $2.5 billion to reach parity. Now multiply this by 36 and you the hill the Google challengers have to climb. Don’t like my $4.00 number. Chop it in half. Still a big hill. The problem is that I have citations that suggest that under certain circumstances Google enjoys a 17X advantage. I don’t include that number in my studies because frankly no one I have met understands this type of gap. Make yourself happy. Pick a number.
Stephen Arnold, April 8, 2009
Android: Surround and Seep
April 6, 2009
I have been following the discussion of Google and its “operating system” since 2002.
My newsreader this fine Monday, April 6, 2009, is chock full of discussions about Google on notebooks, Google on mobile devices, and Google in personal computers. I liked the flurry of comments about the New York Times’s revelation that T-Mobile was going to use Google for its home devices. You can read that bit of news here. Let’s see that means that the GOOG is supporting its own crowd of cronies, Apple’s iPhone, and now T-Mobile. Looks like the Google is getting some traction in the mobile space. Quite a revelation.
Then there is the flurry of write ups about Hewlett Packard’s thinking about Android for its netbooks. The CNet write up is a pretty good one on this topic, You can read Marguerite Reardon’s “HP Considers Google Android for Netbooks” here. Hmm. I wonder if anyone realizes that when installing Chrome, the GOOG has its operating hooks ready for whatever the user wants to do?
In my research for my Google studies, I have quite a bit of contradictory information in my files. On one side of the fence are the Googlers who insist that the company does not have an operating system. The company’s spokespeople are correct. Google has Linux, wrappers, and code shims. The software in use at Google is dynamic and it is not suited for installation by my father on his home computer. On the other hand, Google has figured out how to make Python, JavaScript, and other languages jump through digital hoops. With these software components of which I cannot keep track, Google:
- Delivers its virtual machine technology via containers
- Enables offline access to Gmail and soon other Google Apps
- Creates a digital cocoon in which it can perform such magic as deliver ads regardless of where the user goes or what device he/she uses.
Is this an operating system? It depends. If you are Microsoft, Google’s approach sure seems to be headed that direction. If you are a Googler, this collection of features is little more than extensions of the Google “as is” computing environment–what I call the real Googleplex of one million servers, fancy Dan engineering, and proprietary system sand methods galore.
The point that most of the pundits, mavens, gurus, and Google watchers overlook is the broader strategy the GOOG is using. I dug into this approach in detail in my Google Version 2.0, a deep dive into some of Google’s more current innovations. You can read about that study here. Check out the subtitle too.
In a nutshell, Google’s strategy is to use the Googleplex of its “as is” infrastructure as the wrapper. The Android software to the ill fated Web Accelerator just tap into the mothership. What happens is that Google surrounds the users and competitors and allows its services and features to seep into crack in the existing market sectors. Surround and seep. Quite different from other competitors’ strategies in my opinion.
You can read more about how this works in my forthcoming Google: A Digital Gutenberg study due out later this month. More information about that study is here.
SharePoint Online
April 5, 2009
A happy quack to the reader who sent me a link to Tobias Zimmergren’s SharePoint Online—A First Look” here. On premises SharePoint installations are tar balls that become tar pits. The hapless information technology dinosaurs caught in these traps will struggle and probably die. Uncontrollable costs pull down even the brightest SharePoint wizards in a lousy economic climate.
Mr. Zimmergren’s article makes a very strong case for hosted SharePoint or what the trophy generation consultants call cloud based SharePoint. The idea is solid. Let experts figure out how to make SharePoint behave and maybe perform some useful content related tricks. The users access the needed SharePoint services via a broadband connection.
He does not talk about finding information in the SharePoint system, which is a major weakness of hosted SharePoint. If you can live with the limitations of Microsoft’s approach to indexing, then you are going to be happy. If not, you will have to pursue some other options.
I urge you to read Mr. Zimmergren’s write up. He explains how cloud based SharePoint works and provides useful information to those who may be singing the on premises SharePoint blues.
Stephen Arnold, April 5, 2009
Amazon Embraces Hadoop
April 2, 2009
The fleet footed Amazon surprised me. I read Larry Dignan’s Amazon Launches Hadoop Data Crunching Service” here. What interested me was Amazon’s use of the Hadoop framework. According to Mr. Dignan’s write up,
The service, called Amazon Elastic MapReduce, is designed for businesses, researchers and analysts trying to conduct data intensive number crunching (statement). Hadoop, which is used by companies like Google and Yahoo, is trying to be pushed into the enterprise data center by startups like Cloudera.
I found this interesting for three reasons:
- Amazon has consistently beaten Google to the punch in the cloud computing push for developers and startups. Google has, in my opinion, watched from the sidelines.
- Google influenced the Hadoop system, which is a variant of the Google MapReduce system. You can find a description in my The Google Legacy (2005) here.
- Amazon, despite its early somewhat unusual approach to infrastructure, has gotten its act together. The clearest indication of this is that the company can integrate new technology into its existing data centers and not go down.
In my view, Amazon is making the transition from digital retail operation to a more serious online force.
Stephen Arnold, April 2, 2009
QuePlix: Legacy Data Search
April 2, 2009
Several years ago I listened to a presentation from Index Engines. The company developed an appliance that sat in a back up stream. The idea was that an authorized user could search for a document processed by the back up system. I thought the idea was an interesting one. A number of eDiscovery firms address the legacy data issue via other methods. Today the organization wanting to query legacy information has a number of options.
QuePlix offers a search system for legacy data. Troy Dreir’s “QueSearch: A Search Engine for your Legacy Data” here alerted me to another vendor in this market space. Mr. Dreir wrote:
QuePlix has just released the second of its platform-agnostic programs which are each designed to retrieve information from legacy applications. The first solution was QueWeb, which not only extracts legacy application metadata, but then builds a user interface on top of it. The allows the company to make a transition toward new applications while still using the data from legacy apps. Because it’s based on existing systems, there’s no need to train staff on how to use it and that allows for a smoother migration. The program’s simplicity and usefulness translates into a huge ROI, Tenberg says. QueWeb was launched in 2001 and is already up to its third version.
I did have some information in my files about this company. The key points I had noted when I got a demo in 2007 included:
- The company is a Google partner so there’s an integration capability available to its customers
- Customers can use QuePlix’s cloud option and shift some of the hassles to hosted services such as Amazon’s S3
- A white paper provides more detail. You can get it here.
More information as I locate it.
Stephen Arnold, April 2, 2009
Pentaho: More Open Source Business Intelligence Moxie
March 31, 2009
Business intelligence gets knocks for complexity and performance. In my work, I hear about the direct and indirect costs of business intelligence systems. Those buying BI suites and systems drink the KoolAid and then seemed amazed that the manager of customer support doesn’t know what reports mean or how to get reports that answer specific questions. One of these three weaknesses can be addressed in part by the new Pentaho release. Open source business intelligence lets a customer sidestep some of the licensing fees. Alas, some of the other costs remain.
Pentaho, an open source business intelligence vendor, released a new version of its BI suite. You can read the details here. The company’s tag line for version 3.0 is “User friendly, cloud ready, and community powered.”
This new version includes an updated dashboard design tool. The idea is that an authorized user with no training can create a display of metrics without having to wait in a queue for a developer. The notion of personalization eliminates the grousing that canned reports are of minimal use to most line managers. This new version hooks into Amazon’s cloud services. The design of Version 3.0 makes it easier to integrate specialized functions or “hook in” other enterprise data.
There are basic search tools, but like other business intelligence systems, finding information via a key word query lags the more traditional analytic functions. More information is available here. A demonstration is available, but I had to register my trusty boxer dog as a prospect in order to gain access to the information. I much prefer basic information to be available without requiring me to register an animal. When I know I have a need for more detail, that’s when I would be willing to register. Your mileage may vary, of course.
Stephen Arnold, March 31, 2009
Google Security Method Echoes Microsoft’s No Problem Policy
March 31, 2009
Short take. Google Docs has some security issues. Google’s view of these issues, according to PCWorld here, is to reassure. PCWorld wrote:
In an official blog posting, Jonathan Rochelle, Google Docs’ product manager, details why the company has determined that the issues included in the analyst’s report are far from critical.
Banks don’t talk too much about security. Microsoft doesn’t talk too much about security. Google doesn’t talk to much about security. Factor these sentences and you get, “Don’t worry.” You will have to make your own decision about which services are ready for prime time in the real world.
Stephen Arnold, March 29, 2009
Desktop Operating Systems: A Partial Romp through the Graveyard
March 30, 2009
I was enticed by the title of this ComputerWorld article here: “Gone But Not Forgotten: 10 Operating Systems the World Left Behind” by Matt Lake. I am an old and addled goose, so the amount of detail provided for each of the 10 operating systems varies quite a bit. Mr. Lake does a good job with the highest profiles systems, less well with the older, smaller OSes. What struck me when I read the article was that none of the operating systems differed significantly under the kimono. I grant the coding was different and features available to developers varied. The significant difference was the interface. What I noticed from the screen shots was that the look and feel of the operating systems converged. Over time, the interfaces moved from the inscrutable to the explicit. My take away from the article was that the operating system has become mostly irrelevant to the user. The interface is shifting from the explicit to the anticipative. The implications of this in my opinion translate to significant market upheavals. Who will suffer? Most of the enterprise software vendors will find themselves on the wrong side of shift. Interoperability will eventually become a smart software problem. Products like Chrome, therefore, which look like a browser but are in effect software versions of space ports that connect the world of the user’s data craft with the larger universe, are important.
Stephen Arnold, March 29, 2009
For the Millions of Cloud Computing Ontology Lovers
March 29, 2009
I love ontologies, particularly those created for some of the one day seminars that are available. I love cloud computing analyses, particularly those created by the azure chip consultants who would not walk past, let alone paddle, in the addled goose’s pond. I loved the write up by Kevin Jackson in CloudComputing.com’s “A Tactical Cloud Computing Ontology.” You must read the article here. Mr. Jackson tackled the job of taming two of the most widely used buzzwords at information technology conferences today–cloud computing and ontology.
He presents several diagrams that put the cloud computing idea into a framework. the diagrams are useful, but they do contain some terms that I am not exactly sure how to define. Nevertheless, the distance between cloud computing and its ontology is narrowed. He outlined three actions and considerations the reader may wish to consider. I can’t quote this complete sequence, but I can identify broadly the ideas:
- Merge the cloud and on premises experience so the user doesn’t see much of change
- There will be some differences when using the proposed framework
- Something I don’t fully understand well enough to summarize: “As a way to organize an enterprise’s body of knowledge (architecture) about its activities (processes), people, and things within a defined context and current/future environment.
I am still thinking about how this framework applies to search across secure and open content sources in a regulated environment with known network bottlenecks. My hunch is that others will be thinking about these issues before embarking on a composite architecture. No harm from thinking either.
Stephen Arnold, March 28, 2009
Learning from the Cloud Manifesto
March 29, 2009
I ignored the cloud manifesto, pointing out that secrecy is useful. Obviously the document was not intended to be kept under wraps, so a mini-microblogging storm raged. CNet’s The Wisdom of Crowds ran James Urqhart’s article “Cloud Computing: What We Learned from Manifestogate”. You can read this write up here. The article includes links, an itemized list of the four ways to perceive the cloud manifesto, and a conclusion that strikes a positive note: “open is good.”
In my experience, the clouds owned and operated by commercial enterprises will behave the same way opposing forces have behaved since stone age tribes split into factions and promptly embarked on chatter and warfare. The crazy idea that the cloud operating environments will behave in a way different from other technology battles is off base and not in line with what is now going on among the Apple, Microsoft, and Google camps in mobile services. I am omitting the other players because I don’t want to trot out too many examples, which are legion.
Amazon’s cloud may communicate under circumstances determined by the world’s smartest man who is now working as an order fulfillment clerk about 45 minutes from where I am writing this post. Google will play ball as long as those folks follow the Google rules. Microsoft is going to do what Microsoft has done since its inception and make an effort to enforce its agenda.
Each of these companies will yap about open standards. Each of these companies will put their pet open source wizards on display. Each of these companies will attempt to capture as much of the market as users, competitors and regulators allow.
At some point in the future, the agendas will shift from the cloud to the next big thing. At that point, a big dog will be in the yard and the other dogs will cooperate or get their necks broken. I appreciate Mr. Urqhart’s view. I think we’re in line for a good old fashioned standards battle. Forget cannon fodder. Think column fodder. CNet will be in seventh heaven.
Stephen Arnold, March 29, 2009
-
- Subscribe to Beyond Search
Feature archive
News archive
-
