February 05, 2007
Phil Windley notes that Doug Kaye has posted a description of how he's building a podcasting service using Amazon Web Services' utility infrastructure - not just for storage but for processing and messaging as well. "Dynamic computing power on demand," writes Kaye, in regard to Amazon's Elastic Compute Cloud (EC2). "The dream has been realized." (Kaye needs to work Mechanical Turk into this somehow.)
Zoli Erdos reports on another study showing that large companies are embracing software-as-a-service, not just for CRM but also for content management, project management, and other applications. SaaS, he writes, "is indeed becoming mainstream amongst large corporations: '61% of North American companies with sales over $1 billion plan to adopt one or more SaaS applications over the next year.'"
Commenting on the news that Google will be adding a presentation application to its fledgling office suite, Larry Dignan speculates on "the end game":
In the end traditional Office uses – word processing, spreadsheets etc. – will move online leaving Microsoft's shrink-wrapped Office to become a high-end application. Office 2007 already resembles more of a business intelligence tool than a way to type up memos. If Office moves upstream that'll leave an even playing field for Google and Microsoft to duke it out for Web-based productivity tools. It'll be a protracted war with lots of strategy.
It will be interesting to see whether Google offers a YouTube-like service for easily storing and streaming business presentations online. It amazes me that Microsoft doesn't have such a service up and running for Office 2007.
UPDATE: Business Week reports that Google's business suite, Apps for Your Domain, is about to come out of beta. The package, which has been free to beta testers, will now become a for-fee service, with a monthly subscription charge on a per-user basis. Here's what the magazine says:
After months of dancing around with Web versions of e-mail, group calendars, and the like, Google Inc. is finally about to take a big leap onto Microsoft's turf. Since last August, the search leader has offered a test version of an online office productivity software suite, called Google Apps for Your Domain, that lets companies offload e-mail systems to Google while keeping their own e-mail addresses. Soon, it's expected to add word-processing and spreadsheet services to the suite, which includes an online calendar, chat service, and Web page builder. In coming weeks, Google Apps will turn into a real business as Google begins charging corporations a subscription fee amounting to a few dollars per person per month.
Doug's use of Amazon's service does appear quite non-trivial.
Note, it is Phil Windley (with a "W").
Posted by: Tim Joransen at February 5, 2007 02:56 PM
Typo corrected. Thanks.
Posted by: Nick Carr at February 5, 2007 04:10 PM
Lets not get carried away here.
Leveraging S3 works great for products that require a "one to many" distribution - such podcasts, videos, downloads, image access, photos, etc...
The scale of using cloud storage really breaks down, though, when one tries to incorporate cloud storage as part of their transactional infrastructure (for products and transactions that require "many to many" distribution).
Kaye's description works well since he is also leveraging EC2, but for a company trying to leverage S3 for regular backups or alternate file storage, the bandwidth costs and performance hit of uploading massive amounts of data on a daily basis make it cost prohibitive and inefficient. (Not as a result of AMZN's costs but the vendor's bandwith costs to upload 250GB of data every day)
Adopting or conforming a company's existing infrastructure to S3 sounds nice - and works well for certain tasks - but is not a viable replacement for large swaths of a company's infrastructure.
This tells me that a company that builds an app, from scratch, on EC2 / S3 might have a scale advantage over existing companies - but will be at the mercy of AMZN's cost structure, reliability and up time.
Just something to think about.
Posted by: IsaacGarcia at February 5, 2007 04:40 PM
That "61% of North American companies" quote is taken from the McKinsey & Company study you mentioned here: http://www.roughtype.com/archives/2006/11/cio_interest_in.php. Indeed Zoli's blog links to back to your article. (BTW I still haven't been able to find the original article that you previewed - do you know if its been published yet?)
The "new study" finds that "63% of companies with over 1000 employees adopted some On-Demand solution, vs. 46% of those with less then 1000 employees."
The two studies appear to paint the same picture, but its confusing to read about a new study then have it followed with a (memorable) quote from the old study.
Posted by: Gilbert Pilz at February 5, 2007 07:38 PM
I think that economic (rather than technical) reasons to consider EC2 and similar services are compelling. Recently BNP Paribas signed a multi-million dollar deal with IBM to provide a compute on demand facility. When I did my back-of-the-envelope calculations on that deal (here) it rather looks like going to EC2 would have saved the bank quite a lot of money, with a net increase in the number of avaialable cycles. I rather suspect this would be true for any compute intensive application.
Posted by: James Gardner at February 7, 2007 12:57 AM
Post a comment
"Riveting" -San Francisco Chronicle
"Rewarding" -Financial Times
"Ominously prescient" -Kirkus Reviews
"Riveting stuff" -New York Post