SAF: Day 2


The Architecture Last Mile, Norm Judah

In Microsoft’s view, the last mile is the end-user device, and they believe this is an area that has not received enough attention. Norm made a couple of observations. First, billions of end-user devices are being used and there are many different kinds of devices, ranging from PCs, laptops, cellphones, Xboxes, Zune, home appliances, automobiles, to name a few.

End user adoption and operational excellence are key critical success factors for achieving success in the last mile. To achieve end-user adoption we have to focus on people and their issues. We need to understand the changing demographics and how and where the endpoints are used; aligning information to the personal needs of the individual.

Microsoft has developed a maturity model based on that one developed by CMMI. It defines five stages, starting with: Initial; Management; Defined; Quantitatively managed, and Optimized.

Operational excellence refers to establishing and then managing to SLA’s (service level agreements).

Live is Good, Gary Flake

This was a discussion around Live Labs, a research group within Microsoft, and some of the work they are doing. Albeit interesting, their work is positioned close to basic research that may drive product development some time in the next three to five years.

Gary observed four things happening in the web:

  • Democratization and macroization
    The amateur can provide professional-looking content–the difference between what amateurs and professionals produce is shrinking–as a result of lowering the barriers to individual participation. While this is usually considered in a consumer context, it applies equally to the business world, or more precisely, participation by the smaller business.
  • Power laws and long-tails
    The weight of the long tail is more than the head or more concretely, in combination the sum of all small producers (in long-tail) often outweighs the sum of the large producers.
  • Ecosystems and network effects
    Preparing lists, tagging, mashups of pictures and maps, annotations and reviews are being contributed by those millions of people in the “tail” of the web, increasing valuable content on the web.
  • The innovators dilemma [1, 2]
    The innovator, once established, is unable to move out of their practices; late arrivals learn from the mistakes of the innovator and develop best practices that the innovator is not able to adopt, and in the end looses to the late arrivals.

Gary’ observation is that the internet is becoming a digital mirror of the physical world. This has wide implications, but one implication is that you can operate (test) on the mirror rather than the real world.

Emerging application framework
The discussion tended towards the issues being faced at an application level; the issue that received much discussion was integration: how do you integrate multiple .net clients each developed by a different vendor.

The outcome: software factories seem to be the solution; one can use them as templates to define consistent frameworks.

Web2.0 and the enterprise
As the subject suggests, the focus of this discussion was how does one move Web 2.0 concepts into the enterprise.

The host of the session introduced the term Enterprise 2.0, which had recently been coined to name the convergence of Web 2.0 and the Enterprise. I have included some references below:

For Enterprise 2.0, there are two forms it can take: one for external consumption of enterprise services by customers and one for internal use by employees.

At this point in time, external use seems more viable and applicable to five different problems:

  • product development
  • marketing
  • sales
  • product support
  • training

Basically, it offers a means to keep in touch with the customer. However, one should be aware that the consumer-oriented social networking tools often involve a different technology stack from that typically used within the enterprise.

Internal deployment of Enterprise 2.0 is more problematic. Typically a Social Network requires a very large community as such a low percentage of those will actually contribute (the number suggested was 95% watchers; 5% contributers). The reason social networking is successful on the web is scale: there are millions of people in the community.

Typically, internal deployments are geared towards some form of knowledge management activity. The presenter was dubious about success, having tried it many times only to witness failure. He conceded that success might be strongly related to a pre-existing successful internal social network. However, it was felt that incentives generally play a part (i.e., it is someones job).

The presenter suggested the following references


Leave a Reply

Your email address will not be published. Required fields are marked *