Platform Permavantage

I was listening to Warren Buffet in an interview on CNBC. When asked why he did not invest in technology, he had a very interesting answer — he could not tell who would dominate technology in 10 years. It wasn’t clear if there was any such thing as a long-term competitive advantage.

I thought of a word for this – “permavantage” (permanent advantage, like permalink). Look at the past quarter-century in technlogy (that may be too short a time-frame in Buffet’s world, though). The baton for technology leadership keeps shifting. IBM, Compaq, Microsoft, Cisco, Oracle, Nokia, AOL,Yahoo — everyone has seen leadership come and leadership go. Today, each of these companies faces challenge going ahead. Leave alone visibility for the next decade, many of the companies don’t have visibility for the next quarter!

So, is there such a thing as a permavantage? I think there is, and that comes from owning a platform. Microsoft has succeded in doing that with Windows and Office, though I think it will be hard pressed to maintain that with the next set of PC buyers for whom cost considerations will be of paramount importance. Intel owns the CPU platform, but as the ground shifts to a graphics-intensive world, it too is being threatened by Nvidia. Nokia now has to worry about companies like Samsung because it never really built a platform in cellphones. Brand yes, but not a platform. In fact, all the technology leaders have great brand, but to build a sustainable advantage, what is going to be needed is platform leadership.

That’s what we will need to accomplish in Emergic. We have to think of Emergic as a platform for the technology needs of the world’s emerging markets. Building one or two products may be easy, but building a platform is going to be our greatest challenge. To succeed means getting a hundred things right, and not a single one wrong. Only one company in the past quarter century has managed to come close — Microsoft.

The basic vision of Emergic can be summarised in 3 sentences:
– a computer on every desktop and in every home – for no more than USD 100
– all the software needed for USD 5 per person month
– combine the bottom of the (enterprise) pyramid into emergent networks (where the whole is smarter than the sum of the parts)

Emergic will definitely need to build a brand on the marketing front, and a platform on the technology front. There is an opportunity in the world today — as companies look inward and focus on how to survive the downturn, we need to think disruptively and innovatively to build out Emergic as the Next Technology Platform, covering Computing, Software, Communications and Business Processes.

Enterprise Emulator

While making chips, engineers use a emulator to simulate the chip’s working in software. It would be good to have something similar for enterprise software. What we need is a software system to simulate the enterprise and its ecosystem. It is almost like a video game, a kind-of “Sim Enterpri-City”. A software company can then “inject” its software into the system and see how it performs.

Today, we test software at the sub-system level. But, it is very difficult to get a feel of how it would perform in the enterprise world. When I read about Nvidia testing its chips at emulators from IKos costing USD 4.5 million, it struck me that what we probably need is an “emergent system” of enterprises which can simulate the behaviour of companies and thus create a testbed for new software. Maybe, later, just as video games have gotten closer to reality, this could even be used to test-market ideas and advertising campaigns!

Emergic Update

BlogStreet: We’ve got three things ready in BlogStreet: (i) a Blog Neighbourhood Analyser, which given a blog gives you its friends as analysed through the blogroll and bloglinks on its home page, (ii) a Blog Search which is keyword-based and covers the home pages (iii) Blog Ranking, wherein every blog is given a rank based on its count among the blogrolls. We’ve botted about 1800 blogs. Our program is a little slow, but we are working to optimise it. We should make it available this week. We also have links to a blog’s RSS feed, with the ability to add into an RSS Aggregator. Plan for the coming week: auto-categorisation of blogs.

Digital Dashboard: Last week, we built on the RSS Aggregator to ensure we’d catch most formats. We also added a feature which lists the collective list of blogs being subscribed to by others in the group (enterprise). The key feature which needs to be added now is the Delete option, so it will enable the viewing of only the new entries in the feed (on a per-person basis). Once this is ready, we should (a) make the post option work with Radio and Blogger.com, in addition to MopvableType, and (b) making this available as a hosted web service. The combination of BlogStreet and Digital Dashboard should make it very interesting for bloggers.

Thin Client-Thick Server: Last week, we began our productisation efforts. We are now working towards making an installable set of CDs. This means automating all the steps we had done manually over the past 2 months. We’ve made substantial progress on this front and should have a first-cut CD ready this week. Have also been seeing how the Windows Terminal Server option can be used to provide Windows support for those who may need it (on the Linux Thin Client desktop). We’ve also been testing both hardware and software RAID. This should be the work we’ll be finally moving the Marketing team to Thin Clients.

Enterprise Software: We’ve been doing a study of EJB, J2EE and JBoss (an open source application server). Have also begun a review of the business proces standards – ebXML, RosettaNet and BizTalk.

This Blog: We did a few enhancements: Categories, Search which goes directly to an individual post and GoogleBox. Next up: adding a ListServ so we can send out emails, auto-posting the day’s traffic info to the blog, referer analysis and upgrading to MovableType 2.2. I also created a private blog.

Jabber and Blog Notification

DJ Adams has built on the BlogToaster idea (real-time IM-based notification when one’s favourite blogs are updated):

I put together an experimental bit of code which reflects this mechanism out into Jabber plasma. It’s a pubsub concentrator that sits in front of Simon’s Subscriber Interface and allows any app that can send and receive simple Jabber packets to request and receive weblogs.com-based update pings via this subscriber interface.

This new mechanism abstracts the Subscriber Interface out and allows many subscribers to share one subscription connection. Publish/Subscribe. One publisher, many subscribers. The publisher, in this case the Subscriber Interface, only has to send out one SOAPy ping per updated weblog URL to reach potentially many notification recipients (subscribers).

So rather than reproduce a blogToaster-like mechanism, I thought I’d have a go at putting together a mini-infrastructure on top of which lots of different blogToaster-like mechanisms could be built.

DJ Adams is the author of O’Reilly’s “Programming Jabber” book.

We definitely need to look at this and see how it can be integrated into BlogStreet.

New Search Kid on the Block

WSJ on AlltheWeb: “Norway-based Fast Search Transfer ASA got plenty of attention last month when it announced that AlltheWeb.com , its showcase search engine, had narrowly passed Google in terms of the number of Web pages searched, at 2.096 billion to 2.073 billion. Launched in 1997, Fast also provides the search and filter technology for global portal Lycos, and provides search technology to a range of corporate customers that includes information company Reuters Group PLC, computer giant International Business Machines Corp. and online auctioneer eBay Inc.”

Says its Fast co-founder John M. Lervik: “We don’t intend to be a consumer brand. We want to be the engine under the hood. We intend to provide our technology across all sorts of portals and to businesses. When you start to become a media company, you talk to marketing and public relations types of people, and we want to remain a technology company.”

TECH TALK: Server-based Computing Redux

Here is how Server-based Computing works (“Understanding Thin Client-Server Computing” by Joel Kanter):

All applications and data are deployed, managed and supported on the server. In addition, 100% of the application executes at the server. The application logic is separated from the user interface at the server and transported to the client. This separation means that only screen updates, mouse clicks and keystrokes travel the network.Thus, applications consume just a fraction of the normal network bandwidth usually required. Because applications require fewer resources, they can be extended from one location across any type of connection to any time of client with exceptional performance.

An article in Information Week (June 12, 2000) elaborates on the benefits of Server-based computing:

Maybe the mainframe guys got it right after all. Sure, they didn’t have a cool and intuitive graphical desktop like the PCs that are sprinkled throughout the enterprise, but they did come up with an architecture that provided the benefits of centralized management, a relatively low cost of ownership, and a resource that was clearly made for conducting business.

Visualize a centralized architecture combined with the power of the graphical interface and a simple desktop or even handheld user device. Think about simplifying the desktop, migrating complexity into a centralized controlled environment and minimizing weak links in the enterprise. Here’s what Gartner Group, Tolly Group, and Zona Research tell us server-based computing:

  • Empowers companies to reclaim 57% to 80% of costs associated with enterprise desktop computing.
  • Decreases staffing requirements by enabling administrators to manage 500-plus users, primarily through the concentration of up to 200 users per application server.
  • Reduces liability at the desktop and creates a zero-desktop-administration situation.
  • Introduces time-to-value benefits for deployments and software version control.
  • Improves performance of centralized enterprise apps, particularly those delivered over WAN and low-bandwidth connections.
  • Provides gateway connectivity for supporting heterogeneous systems on the back end while fostering a thin-client device on the desktop.
  • Provides the company with an architecture well-suited for long-term expansion, and flexibility without extensive system changes.

  • A more recent article by Andrej Volchkov in IT Pro (March-April 2002) weighs the pros and cons of server-based computing:

    The principle advantages [of server-based computing] are reduced maintenance and support for client terminals, a standardized corporate client terminal and centralized resources management. All of these in turn make it easier to support increased employee mobility and the deployment of heavy client-server applications at remote locations or subsidiaries because of less bandwidth usage.

    Server-based computing has disadvantages also; among the most important are that the terminal server represents a single point of failure. It is also not very suitable for deploying small business applications requiring local country or language support. Applications (particularly graphic-intensive ones) can also experience performance problems or become unavailable because of network problems, heavy printing demands, and large file transfers over low-bandwidth connections.

    Even as we consider the benefits of server-based computing, we cannot overlook the downside the central server can be a single point of failure. Though technology has become increasingly reliable, it may become necessary to deploy a bank of servers which are load-balanced, rather than a single server. In addition, storage can be separated from the server through the use of a NAS (network-attached storage) or SAN (storage area network). It depends on the end user segment one is addressing and how much the customer is willing to spend. In fact, the most inexpensive solution could consist of using software-RAID on a single desktop computer which is used as a server, and using diskless desktops with limited processing power.

    Tomorrow: Citrixs Solutions