Personally, I spent quite some time on the Enterprise Software Re-thinking. This is part of the Tech Talk series which has been ongoing for the past 3 weeks, and will continue for the next 2 also. The pieces are now coming together quite nicely. The more I think about it, the more I believe that the approach we are taking is the right one. The areas which need more work are (a) how to build the platform for developing enterprise software components, and (b) how are we going to market and sell to SMEs.

BlogStreet: Our proglets (10 of them) are working to identify blog rolls. They work reasonably well. We now need to put together a Blog Neighbourhood tool on BlogStreet. This week, we'll be thinking through on how best BlogStreet should be -- as a resource for bloggers and readers (there are 1000x more readers than writers).

Digital Dashboard: We were stuck for a couple days on a problem with the Apache-SOAP interaction, but finally fixed the problem. Have an RSS Aggregator in place now which uses the Outliner web service we built. Now, we are working to put the blog interface together, so we can get everyone internally creating blogs. Blogs+Outliners+RSS=Digital Dashboard.

Thin Client-Thick Server: Except the Marketing department, everyone's now using Thin Clients. We bought a few more second-hand PCs, each for Rs 7,000 or so (including monitor). There have been some issues with the stability and hardware -- rising load, increasing processes, disappearning icons, crashing applications. Nothing serious -- just the occasional hiccups. Our target is that by the month-end we have a completely stable environment. Seeing how things are going here convinces me we are on the right track.

Enterprise Software: Got stuck with some problems using SAPDB. Decided to then drop it. Focus now is on using Postgres, JDBC, JBoss and J2EEas the building blocks. All are new areas for us.

Messaging: Have worked out a set of ideas to strengthen the core platform, and also thinking on how to integrate the Mail Server with the Thick Server. We also are thinking on a plan to expand in multiple cities.

This is the fun part of being an entrepreneur...each day brings its own share of unknowns. Thats why I strongly believe that for an entrepreneur its not just reaching the destination which is important, but the journey itself should be fun -- one has to like, in the words of Dan Bricklin, "jumping from rock to slippery rock."

I love Writing. Its been a habit formed early on in life. I used to keep daily diaries when I was in college and IIT. This continued for some time in the US and later in India, but then as life got busy and more challenging, the diaries took a backseat and then stopped altogether. I didn't like writing depressing things! Even though the diary habit hasn't come back altogether, I do occasionally write my thoughts in my notebook (paper and pen, not electronic).

Over the past few years, my note-taking habit has become very good and disciplined. Earlier, I used to write on pads and whatever book/diary I'd get. A few years ago, I got the idea of using 300-page notebooks -- this came when I watched a person in a meeting do so. The good thing about using big (in terms of pages) notebooks is that one doesn't try and economise to conserve pages! I now write (meeting notes, doodling, thinking) 10-15 pages a day. This has been a great asset for my thinking. Even when reading, I make notes of the key points -- helps me remember better.

Then, there's been the public writing. In 1995-96, it was a fortnightly column of 1500 words for Express Computer. I stopped because there wasn't much new happening then and I found myself repeating myself. In 1997-99, I wrote occasionally for IndiaLine, our own website. Wasn't very regular. Late 2000, I realised that my thinking had dried up and I wasn't as much aware of the world of technology as I used to be.

That's when I took the decision to start a daily Tech column on Tech Samachar. I began by writing the first 3 weeks worth of columns (15 in all, about 400-500 words each). I wasn't sure how long it would last. Now, more than 18 months later, I am myself surprised! This writing has not only broadened and deepened by writing, but also helped consolidate my thoughts. I write what I think. I write because I now love it. It has made me much more alert, much more thoughtful, and more able to connect things together.

The Weblog is now an extension of this writing. It has now become part of my daily routine. It is also working as an extension of my memory -- when I come across something interesting, I post it to the blog.

There is a significant time investment in the writing. Each week's Tech Talk series of 5 takes me an aggregate of about 3 hours. Each day, I spend about 30-45 mins on the weblog, for a total of about 4 additional hours each week. But I think the time investment is absolutely worth it. This is what gives me the new ideas, it has introduced me to new people and it gives a purpose to my reading. Its also my way of giving a little back to the community of writers (professional and amateur) who have been the inspiration for me.

The central thinking behind the enterprise software objects is that of events and (near) real-time computing. This is the philosophy outlined by Vivek Ranadive of Tibco in his book “The Power of Now”. The concepts are many years old, but they haven't been applied to Small and Medium Enterprises (SMEs). Software for SMEs is typically tightly-coupled and biased towards batch-processing. Real-time computing is what needs to be delivered to SMEs.

We need to think of (near) real-time instead of real-time because, for SMEs, the cost of setting up a real-time, fully connected network across multiple locations can be extremely high. However, if we are willing to tolerate a slight latency, then the costs can be brought down dramatically through the use of the Internet and the “intermittent connectivity” that is available in most emerging markets.

Says Kenamea CEO John Blair (as quoted in Release 1.0, December 2001): “If the browser enabled the first-generation Internet, and the application server enabled the second, then Internet-scale real-time communications will enable the third – true network applications that are dynamic and fully interactive”.

This is the leap that SMEs need to make: most are stuck in the first-generation of the Internet. They need to leapfrog, and this is where software components and standards will help them.

What are the key components of the new enterprise IT architecture? The software part consists of the Enterprise Core (the Enterprise Processing Unit), the Enterprise OS, the Information Bus, the adaptors to connect to existing legacy applications and the Corporate Portal.

Enterprise Core: This consists of the system Operating System (Linux) and related components (file server, print server, web server, print server, etc.) This also has the notion of users and groups. It includes other building blocks like the Mail Server and the Instant Messaging Server. It also supports real-time alerts via SMS to cellphones thus ensuring a real-time messaging and notification capability.

Enterprise OS: The Enterprise OS consists of the basic enterprise building blocks – the set of generic applications and objects which are common across enterprises. For example, it is likely that 60-70% of accounting fundaments are exactly the same across organisations and even countries. These get embedded into the Enterprise OS. The same applies to the other verticals also. So, take the common concepts across verticals and integrate them together into the Enterprise OS. The Application Server is the container for these objects. Examples of Application Servers are Jboss (open source) and BEA Weblogic and IBM Websphere.

Information Bus: The building block for the real-time enterprise is the Information Bus. What this means is that as events are generated, they get published on the information bus, and “subscribers” take the events and process them appropriately (based on the rules). Concomitant with the information bus, events and publish-subscribe are two additional concepts: subject-based addressing and IP multicast. Subject-based addressing ensures that events are not addressed to specific recipients, rather the filtering happens by receivers based on the subject field and what they have subscribed to.

An alternative to publish-subscribe is request-reply, which is what we are more familiar with. When we go to a website and request info for our bank account balance, that is in request-reply mode. Publish-Subscribe would mean setting up rules to be alerted in the case of certain events happening (for example, the bank balance falls below a certain limit). Publish-subscribe enables “management by exception” -- where rules can be encoded in business processes to only alert if certain events happen.

Thus, the Information Bus is the backbone on which the events, messages and data flows. It inter-connects all the modules. It supports the Web Services API, allowing the modules to have their functionalities accessible by other modules. Commercial implementations of information buses include Tibco and IBM’s MQ Series. Newer Internet-based alternatives to real-time messaging buses come from companies like KnowNow.

Databases: This is the storage system – like a computer’s Main Memory. What is needed is for the database to provide an integrated view of the data.

Adaptors: Adaptors provide linkages to existing applications. They bridge the apps to the Bus. For example, adaptors can provide linkages to data in existing applications like Tally (accounting) or even MS-Excel or databases.

Digital Dashboard: This is the unified view of the enterprise. It becomes the new desktop. We will be discussing this in more detail later.

Tomorrow: Software Components (continued)