« Mainframe = Youth | The Mainframe Blog Home | Elephant Man »
Google bringing back the mainframe to render Microsoft obsolete
I thought that headline might get your attention.
Certainly such a notion seems like one we should consider on mainframe blog. For those of you that don't know, Om Malik is an astute, and very widely read, industry watcher. Anyhow he gave some of his blog real estate to a chap called Robert Young, to make an argument about where Google is heading, and what this direction means for Microsoft.
So what is the quote that caught my eye?
To some extent, Google is bringing back the architecture of the mainframe to render Microsoft obsolete. In the future, all computing devices, whether it be the PC, mobile phone, TV, etc., will simply be terminals that “plug-in” to Google’s massive server grid and application services. With the increasing price/performance of CPUs, memory, bandwidth, and storage, Google’s strategic edge will be based on their advantageous cost of processing bits.
The advantageous cost of processing bits - lovely phrase that. Sounds like a mainframe mantra. The kind of thing Peter Armstrong might say.
Of course Google doesn't actually use mainframes to achieve its incredible response times and data management capability. But then again, it doesn't need to do a lot of things mainframes do. There is no such thing as the transactional integrity of a web search.
Google is very much a scale out architecture, based on clusters of low cost machines, rather than a scale up, slice and dice architecture. In that sense Young is somewhat off track when he makes the architectural comparison.
But the parallel is clear in the sense of a shared network resource which thin clients (browsers) access to manage information needs. Google increasingly is our memory, just as the mainframe for a long time was the undisputed corporate memory. Note also how much of Google's success so far is text-based; its revenue stream is based on text-based adwords, and the google interface is incredibly simple. It might as well be a 3270 terminal.
What are the lessons for mainframe communities, developers and evangelists?
Well for one, don't worry about fancy client widgets and so on. If the business asks for a new fancy interface for a mainframe application you can ask why. Google is a great argument for the power of simplicity and constraint-based development.
How many of the features in Word or Lotus Notes do you actually use?
One of the important movements in IT at the movement is a drive to simplicity. Value lies in embracing constraints. The poster child for this approach is 37signals, a really interesting ISV that actively fights against adding every damned feature anyone ever thought of. They have a coherent philosophy and uniquely creative approach. The 37signals blog signal vs noise should be essential reading for anyone considering the trade-offs between new features and functions, and their actual value to end-users.
The argument is not that design is unimportant. On the contrary good UI and application design is essential to effective systems development.
Some mainframers might ask what does all this have to do with me? I would argue keeping abreast of new approaches, terminologies and trends, new languages for describing problems and how to solve them, will enable you to have more fruitful discussions with business owners and other technology groups.
Could you turn your mainframe into your corporate Google, taking advantage of all system of record stuff? Why should Microsoft, or even Java, be an essential element in displaying and presenting mainframe data? What data services could you present that would excite end users, given that Information is often far more important than cool whizzy features. Could you index and cache some relevant data, so queries don't chew up cycles, and offer this up as a service for consumption for application developers in other parts of the organization? Start driving grass roots service oriented development.
The mainframe is certainly graphically challenged but that is not necessarily a bad thing; its just a constraint. Most blogs aren't about whizzbang graphics, they are about text. So too wikis. XML is text.
Microsoft is not going to become obsolete, any more than the mainframe has, but it certainly makes for a good headline, and hopefully some bits for thought.
| by James Governor | August 26, 2005 in Innovation Permalink |
TrackBack
TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00d834521c8469e200e55069b4998833
Listed below are links to weblogs that reference Google bringing back the mainframe to render Microsoft obsolete:
» Long live the mainframe! from Daniel N. Meredith :: kodiak
The Mainframe blog has some great thoughts on the that the world is slowly coming back to. Google (or any large data grid) is kind of a mainframe if you squint at it and tilt your head. Sort of an... [Read More]
Tracked on Aug 30, 2005 10:03:54 PM
» Google competes with Apple, not Microsoft from James Governor's MonkChips
So apparently IBM keeps Gates up at night. Not still having nightmares about the 360 coming back to show you how to integrate a database with an operating system are you, Bill?Well just as a contrarian brief, I have been... [Read More]
Tracked on Jan 5, 2006 1:38:05 PM