Search Architecture Meet a Nemesis
March 24, 2009
Hinchcliffe.org had an intriguing article called “10 Must-Know Topics For Software Architects In 2009”. You can read it here. I thought the identification of the “must know topics” was good. As I ran through the list (cloud computing, non relational databases, etc.) I thought about the implications of this list for search and content processing. Boiling down my mental flapping I came up with one sentence:
Search vendors and software architects are on different wave lengths.
None of the 10 points are part of the brand name enterprise vendors’ search solutions. I can identify some notable companies unable to offer more than key word search. So much for “new application models”. I know that social content befuddles some seven figure systems. The notion of Web oriented architecture is easy for some search vendors to talk about and quite a bit of work to implement.
What’s this suggest? Three points to me. Feel free to lash me with your own notions:
- Traditional search companies are highly vulnerable to solutions that focus on one or more of these key points, embed content processing, and package the fix as a higher value than a point content solution.
- The top 10 challenges are going to be disruptive due to their cost and the pressure on new hires to get access to the features of the “new” architectures that 20 somethings covet.
- Security will become an even bigger problem than it is today.
Do search vendors care? Some do. Some don’t. Do software architects think about search vendors? Not too often.
Mr. Hinchcliffe’s analysis provides some insight into future excitement in the enterprise.
Stephen Arnold, March 24, 2009
Microsoft Load Balancing: All or Nothing
March 24, 2009
I found this technical tip quite interesting. I am not a fan of either – or approaches. Your mileage may vary. Click here to read “Windows Network Load Balancing – Don’t run with the defaults!” from the At Scale Web log. The topic is an important one, load balancing. The tip is to disable the affinity mode, which is the default. At Scale recommends that we “set every single one of your servers in the NLB cluster to non-affinity mode and you are golden.” Either – or. Let me know if it works. I prefer a different architecture with good old dedicated load balancers, but I am an addled goose with little tolerance for some of the Microsoft software load balancing excitement.
Stephen Arnold, March 25, 2009
SharePoint Incremental Update Slow Down
March 23, 2009
Microsoft SharePoint is an interesting system. Originally a content management lash up, the product now offers a wide range of features. There are various flavors of search and dozens of vendors who offer snap in components to address shortcomings in the core product. If you have a vanilla SharePoint, you will want to scan “SharePoint Incremental Crawl Taking a Long Time after Adding or Removing a User” here. There is no direct fix for incremental indexing slow downs when a SharePoint installation has user additions or deletes. If a reader knows of a fix for this issue, please, use the comments section of this Web log to post the remediating method. Interesting issue which strikes me as a significant problem in organizations where new users come and go in a SharePoint centric organization.
Stephen Arnold, March 23, 2009
Oracle: Throwing Hardware at Performance Issues
March 23, 2009
In my limited experience, the easiest way to resolve Oracle database performance issues is to throw hardware at the problem. The problem is that adding hardware can pump up Oracle license fees. The resources needed to engineer non hardware solutions to Oracle database performance problems are often unknown and therefore hard to budget. So, I throw hardware at the problem. Oracle and Hewlett Packard have joined up to make it somewhat easier to throw hardware at an Oracle performance problem. You can read about the new range of slim line database machines here. Computer World reported that “The scaled-down version has four HP DL360 database servers, two switches and seven storage servers and ‘will provide similar scalability and performance characteristics as the full rack.’” The idea is that the slim appliance or appliance-like rack will cost less than the $5 million charged for the tubbier version of the database machine. Oracle database administrators know how to get their share of available information technology budgets with hardware procurements of this magnitude. As the economic noose tightens for more organizations, will Oracle and HP face push back in the months ahead? In my opinion, yes.
Stephen Arnold, March 23, 2009
Salesforce.com and Twitter
March 23, 2009
I found this Network World article suggestive. “Salesforce integrates Service Cloud with Twitter” here reported that the cloud computing CRM vendor has hooked Twitter into its system. On the surface, this is interesting because Salesforce.com sees value in what folks in Harrod’s Creek mostly don’t understand. Network World reported:
Salesforce.com showed how someone working for a telecom provider could spot and track a discussion about a headset that a user was having trouble with. The telecom worker could dig through their company’s internal knowledge base and then send the Twitter user a link to a help document.
My thought upon reading this article was that Salesforce.com is one of the company’s that some Googley engineers watch. I think that the Salesforce.com experiment may provide the GOOG with useful information about the application and utility of real time search.
Twitter in its present form is not flawless. The Salesforce.com beta may provide solid data about what must be done to hook real time search into commercial systems. Twitter may not be the long term winner, but it is a great test bed in my opinion.
Stephen Arnold, March 23, 2009
Big Media to Google, Make Us Number One Again
March 23, 2009
I enjoyed Steve Rubel’s article “Media Companies Ask Google to Favor Their Content Over Blogs” here. He presented the argument originally set forth in Ad Age and some some useful comments. For me, the most interesting was:
A neutral Google is a good Google. They should continue to deliver an algorithm that rewards the highest quality sources that have earned a following, interest and links from other sources. If the media companies don’t want Google to favor bloggers, why not just stop linking to them or use no follow tag? That may over time, erode their Google Juice. However, I suspect most realize it’s too late to put the genie back in the bottle.
Neutral? Hmmm.
However, I wanted to ask several questions. I don’t want to forget them:
- Why are the big media outfits so confident that their information should be at the top of a Google results list? When I run a query about Google or Microsoft technology, I skip big media write ups and look for solid information in technical papers or from specialist sources.
- What’s driving this proposal at this time? My hunch is that after a decade of ignoring Google and even longer hoping that online would behave like information on paper or on 1950s broadcast TV, the big media folks realize that they are marginalized. The savior is Google. Google is not a religion, so why not pay Google for placement?
- How can informed people perceive Google as objective?: Run a query for enterprise search. Who is at the top of the results list? Why is this entry at the top of the results list? Why are pointers to my Google patent search buried in the Google search results? I must admit that the notion that Google is objective is a novel one to me.
I liked Mr. Rubel’s analysis for the most part. I think his write up will spark a number of comments.
Stephen Arnold, March 23, 2009
The Guardian’s Observer Sees Trickiness in Google
March 23, 2009
I am an addled goose and my prose does not flow trippingly on the tongue. I do enjoy British humor. If I did not, I would not have been able to sustain my 25 year friendship with my British publisher.
If you enjoy the Oxbridge way, you will find some enjoyment in Robert McCrum’s “Is Google Committing Theft – or Ushering in a Bright New Age?” here.
Poor Googzilla. Its alleged copyright transgressions continue to provide ammunition to its critics. But Google gets off with a slap on its snout. Mr. McCrum targets “a pop academic” named James Boyle. He was, if I am understanding Mr. McCrum’s argument, used as an example of “a rallying cry to the Googletariat: ‘Nerds of the world unite, you have nothing to lose but your urls.’
I quite like the “nerds of the world” phrase. I don’t care too much for Googletariat. I hope Mr. Boyle experiences joy when he realizes that his book The Public Domain is the focus of Mr. McCrum’s wit.
The only issue I have with Mr. McCrum’s write up is that he works for a dead tree outfit that may face some challenges that wit cannot resolve. I suppose the financial mavens at the Guardian / Observer could use Google to look for some ideas. Well, maybe not? A manual search of the shelves in the basement of Blackwell’s in Oxford would probably be more comfortable.
Stephen Arnold, March 22, 2009
Intelligenx Powers Mapaspublicar.com
March 23, 2009
Intelligenx and Publicar SA recently launched http://www.mapaspublicar.com, a new local search application that focuses on the country of Colombia. Publicar, a leading multimedia content provider in Latin America, is using Intelligenx’s search platform with advanced mapping features and Google API to serve up 3,000 points of interest (e.g., museums, hospitals, schools) and detailed information about businesses in Colombia. The new web site functions like Google Maps to search for Colombia locations, all through a Spanish language interface. It’s another step by Intelligenx, http://www.intelligenx.com, which has more than 10 years’ experience in local search advertising, to create business with more directory publishers and local search providers. The plan is to expand coverage of mapaspublicar.com to eight countries in Latin America.
Jessica W. Bratcher, March 22, 2009
Apple Google: The Arrogant to Teach the Arrogant
March 23, 2009
I found the CNet article “What Google Should Learn from Apple” here amusing. In my opinion, some of the CNet articles are he-said she-saids. Mr. Matyszczyk’s approach appeals to me. He used a Google arts and crafts person’s resignation as a spring board to this idea: Data crushes artistic instinct. I think he’s was right. For me, the most interesting observation in the article was:
The fact is that human beings are astoundingly, depressingly, maddeningly human. Which makes them irrational, contradictory, capricious and, sometimes, just plain nuts. These aspects are the hardest for engineers to get their talents around because, one hopes, they are impossible for engineers to get their talents around.
I have just one thought which I wish to capture. Is this the arrogant leading the arrogant?
Stephen Arnold, March 22, 2009
Newssift Technical Plumbing
March 23, 2009
Thanks to the readers who sent me information about the new “test version” of the Financial Times’s news service. I hope it revives the Financial Times as an online financial news source. I know that the FT has a solid brand and great potential.
Some of the information about vendors pointed back to TechCrunch; other readers just made statements which I will pass along for additional comment / correction. Here’s the line up:
- Endeca–the guided navigation company
- Nstein–content management (started life as a content processing company but changed and now reports record revenues)
- Lexalytics–the new entity formed with the merger / fusion of Lexalytics (sentiment analysis) and Infonic (information management)
- ReelTwo–search, data analysis, and “custom portals”.
My take on this use of multiple technologies:
First, the Financial Times’s beta makes clear that no single search and content processing system can meet the needs of a client like the Financial Times.
Second, the Financial Times implemented a try try try strategy before taking a clear sheet of paper and figuring out how to make its content more accessible to its target user group. I don’t think I can estimate the cost of the present system because it makes clear that earlier efforts at search failed. Those “sunk” and “opportunity” costs are wiped away, but a full accounting of the total cost of making FT information available to its users is more than today’s chief financial officer wants to put on his / her books for an ROI calculation. The same multi year investment in search plagues another European publishing company as well. The problem is not unique to the FT, and that’s important. The shift from traditional publishing business models and methods to Internet models is neither easy nor obvious.
The main FT.com splash page with a welcome screen that obscures the news I wanted to view.
Third, who is the intended user? The site offers a number of powerful functions. The folks who want these types of online operations may already have them available without charge from such places as http://finance.google.com, http://finance.yahoo.com, or (hold your breath) American Online here. As a side note, the AOL service (linked to via Google Finance) runs on the potent Relegence platform here.
The America Online splash page for business and financial information. Note: the information is not obscured by a pop up greeting. Remember. This is the deeply challenged America Online and it is handling business information with its own technology, not a collection of four discrete systems.
Fourth, I think the FT is late to the party. Maybe too late? The company has knuckled down to create Newssift, but the window of time for making big traffic gains has closed. Financial information and analytic tools are available to investors with online brokers such as Fidelity and TDWaterhouse. Business news is available from high traffic outfits like Yahoo News and lower profile services such as Newsflashr.
Fifth, one wonders if the price tag for integrating the various technologies has been tallied. What happens when one of the three or four vendors makes a change? I don’t have sufficient data to estimate these costs. Perhaps the costs are trivial? Somehow I doubt it.
In short, the FT is trying again. Like other companies shifting from dead tree business models to the crunchier online variety of business model, the timing is not optimal. I wish the FT and its vendors good luck and fair weather. My weather charts predict stormy seas ahead followed by a flood of red ink rushing from different points on the compass.
Stephen Arnold, March 23, 2009
-
- Subscribe to Beyond Search
Feature archive
News archive
-
