Grid Computing

From eServer Magazine, a story on Grid Computing and what IBM is doing:

Rather than countless PCs sharing tasks, imagine a large cluster of mainframes and other servers acting as a single integrated computing system that can be accessed by home- or business-based PCs. Beyond simply sharing files and information, users can access an extensive grid of high-end enterprise computers via the Internet, using idle compute cycles to crunch massive amounts of data. Certified users could access super-computing resources, computing clusters, data-storage systems, data sources and even network with other grid users. Taking it even further, users could create entire virtual organizations with complex business processes running remotely. The only costs involved are likely to be those charged to lease grid resources.

Middle East Tech – Dan Gillmor

Dan Gillmor’s conversation with conversation with Yossi Vardi, Israel’s most famous technologist. Vardi created ICQ. He says: “For something to provide a wonderful user experience, you need a touch of . A great user experience has to do with some kind of rarity or uniqueness.”

From another report by Gillmor on Jordan:

Jordanian leaders want to embed technology into the fabric of education and the economy, to help provide essential skills and tools to the emerging generation of adults in a nation where the majority of people are under the age of 18.

The ambitious goals include wiring up every school and university in Jordan with true broadband Internet access. By the end of 2004, leaders hope to have spurred the creation of 30,000 new jobs, $150 million in new foreign direct investment and annual technology exports of $550 million. This might sound modest elsewhere, but for a nation with a per capita annual income of under $2,000, it would be an impressive achievement.

This is the kind of vision emerging markets need – how to get a computer on every desktop and in every home. This investment and enablement via technology is, according to me, the primary requirement for progress.


From the NewYork Times,an article on the future direction of supercomputing — At Los Alamos, Two Visions of Supercomputing:

By 2010, scientists predict, a single chip may hold more than a billion transistors, shedding 1,000 watts of thermal energy – far more heat per square inch than a nuclear reactor.

The comparison seems particularly apt at Los Alamos National Laboratory in northern New Mexico, which has two powerful new computers, Q and Green Destiny. Both achieve high calculating speeds by yoking together webs of commercially available processors. But while the energy-voracious Q was designed to be as fast as possible, Green Destiny was built for efficiency. Side by side, they exemplify two very different visions of the future of supercomputing.

Mercury Interactive

Yet another success story from Israel. A storyinthe New York Times — Roots in Israel, Head in Silicon Valley:

Mercury Interactive, a Silicon Valley company with roots in Israel and offices in 25 countries, soon became a force in the new field of Web management. Its client list now includes about 75 percent of the nation’s 500 largest companies; government institutions like the State Department, Army, Navy and Air Force; CBS’s; and the author Stephen King. While software testing provides the bulk of Mercury’s business, Web management has grown to produce more than 20 percent of revenue in just over two years.

Now Mercury has repositioned itself again, having figured out a way to profit from the dot-com implosion by helping businesses tune information technology they already own.

Milestone: Posting from RSS Aggregator to Blog

Just achieved a breakthrough with my previous post. It was done via our RSS Aggregator directly to this weblog through the MovableType API (actually, the Blogger API). This is what I’ve been wanting to do for long! Be able to scan a lot of entries in the RSS Aggregator (having subscribed to many feeds) and then a 1-click (well, almost) post directly to my blog. No cutting and pasting URLs and titles.

Our RSS Aggregator still needs some work, but we’ll clean that up in the next few days. The part before the RSS Aggregator is also interesting – we are maing a blog directory and neighbourhood analyser (using blogrolls and bloglinks). So, when I come across a blog I like, I can do a 1-click addition of that feed directly into my personal RSS Aggregator. (Currently, it shows the RSS feed link for various blogs, but we’ll get the rest of it working soon.)

This is the flow I’ve been waiting for [see: Blog Enhancements].

Mozilla opens up Microsoft’s closed Outlook PST format – Udell

Writes Jon Udell: “In no time flat I had my Outlook mail sitting in MBOX (i.e., plain text) files under Mozilla’s mail tree. Importing my Outlook contacts was equally successful…As a bonus, I tried again to establish signing/encryption capability in Mozilla’s mail client, something I’d failed to achieve with the pre-1.0 code. Got that working too. For the moment, since Outlook is a better PIM than Mozilla, I’ll probably continue using it. But it’s great to know that Mozilla’s mail client can now work with my Thawte freemail certificate. And knowing that it can also readily force Outlook to regurgitate mail and contacts as usable text files really sets my mind at ease.”

Points to keep in mind for us, because we are going to face this challenge in enterprises if people decide to migrate from Windows to Linux. One of the biggest barriers is Mail which is in Outlook Express or Outlook. Looks like this is a barrier no more.

Rising China – NYT

Writes the New York Times in an article entitled China Races to Replace U.S. as Economic Power in Asia: “As it buys up goods, parts and raw materials from its neighbors as never before, China has accompanied its new heft with diplomatic efforts to assure them that it wants to offer cooperation, not competition. Many have rushed to China’s embrace and are nimbly shifting their economic alliances, particularly as the United States makes its way through only a tentative economic recovery….Mr. Lardy says China’s trade is now growing at a faster rate than Japan’s growth during its boom years in the 1960’s and 1970’s.”

Why Desktops are still King

Fortune compares desktops with laptops and concludes that “as long as your usage pattern resembles the norm, the lowest-cost, highest-performance option for your company is a desktop.”

I feel the same way about desktops. That’s the thinking behind focusing on using desktops and not PDAs or other small devices as part of the effort to bring computing to the mass markets. People need the power of a computer. The only difference is that much of the world can, with intelligent use of software, use 3-4 year old desktops equally effectively, and bring down cost of ownership by 70-80%.

That is our thinking in the Thin Client-Thick Server, with second-hand PCs serving as the desktops.

Alsop on ASPs

Alsop’s contrarian view on ASPs: “There are a couple of ASP-style companies now doing good business. Their secret is that they don’t rent services that are core to corporate IT; they don’t try to get companies to trust the heart of their computing to someone else. Instead, they have shown IT managers that they can do a terrific job running discrete applications. In such cases companies may well prefer to rent rather than license software….The two providers that I’m thinking of are Webex and”

Our approach with the Thick Server is that of an “distributed ASP” — an ASP on the LAN instead of the Web. This way, we get past the connectivity issues so prevalent in emerging markets and the “I-want-my-data-in-my-office” desire.

We are starting by focusing on the core IT infrastructure with the Thin Client-Thick Server approach. But our target segment is one that hasn’t been exposed much to computing. The discrete (and integrated) business applications will come over time.

Opera’s CEO on the Browser Market

Says Jon von Tetzchner:

You’re seeing less changes on the top, but obviously there are changes taking place underneath…[The browser] will have to have voice. But the most important thing is that the marketplace will change. Browsers will be available on other kinds of devices. People have been waiting for this for some time, and everyone agrees that it’s just a question of time. Mobile phones, for example, now have WAP (Wireless Application Protocol) browsers. Over time, we will see a full browser on mobile phones with more advanced screens and hardware. You will also see more set-top boxes and TVs using the Web (browser). Cars, too.

The browser will be there, but what’s needed is a kind-of Digital Dashboard on top of it. So, when I login, it automatically open the browser on my desktop with multiple tabs to my new read-write environment built with RSS Aggregators and Blogs. Going to sites using browsers is what we do now. In future, I think we will be using browsers more as an interface to navigate information that comes to us.

Intuit pushes at Small Businesses

One company to watch and learn from is Intuit, It is exclusively focused on targeting small companies. This year, its been making acquisitions of other cos. targeted at small businesses, the latest one being Eclispe. [ story].

Here’s a little older story providing some background, also from

J.P. Morgan analyst Adam Holt said there’s still a lot of potential in selling products tailored for specific business segments, as the main competition for [Intuit’s] QuickBooks is smaller companies that make software and have only niche products. By coming up with its own approach and selectively acquiring companies such as American Fundware, a specialist in accounting software for nonprofits that Intuit agreed to acquire earlier this month, Intuit can offer customers the best of both worlds.

“It’s a highly fragmented marketplace,” Holt said. “You’ve got a lot of little solutions-providers targeted at individual vertical markets, but there isn’t anybody with Intuit’s brand name, balance sheet and installed base that’s rolling out vertical products. I think they’re really well-positioned to succeed there.”

Intuit is also working to use the strength of QuickBooks to sell other services, such as payroll processing.

We will also need one “killer platform” to build on. Can that be the Thin Client-Thick Server platform? Or the Digital Dashboard? Or Messaging?


Economist on Auctions: “For all their promise, many auctions have strangely disappointed. For online auctioneers, a problem is that customers face costs in time spent bidding, and the risk of losing the auction. Fixed prices can work better if they reduce time spent buying and the risk of not getting what you want. This may explain why few auction websites – eBay is the notable exception – have prospered.”

Economist Survey of America

In a section entitled “Integrator-in-Chief”, the Economist writes:

It is a truism, of course, that if America does bad things or makes bad mistakes, others will criticise it, shun it or even oppose it. It is best, however, to think about this emerging issue of American power and leadership in three ways:

– America’s power relative to its rivals’ and to other alternatives, including its allies.
– America’s power relative to the challenges it faces around the world, and what it might achieve by using it.
– America’s power relative to its own willingness to use it, or to keep bearing the costs of maintaining and using it.

This survey will explore those topics, and the questions they raise. Its answers – be warned – will be optimistic, and will generally be favourable to the United States. Certainly, American leadership will produce mistakes. But without American leadership, worse things would happen. And if any other country were in the lead, there would be much greater cause for worry.

Re-read the above replacing America and US with Microsoft and country with company.

Software Licencing – WSJ

An article on the legal issues surrounding the sale and licencing of software: “We’re just making a one-time royalty payment for a perpetual license to use this software. You see, each program in this box is distributed under a license agreement. According to that agreement, we’re not buying copies of the software, we’re just getting a license to use it, subject, of course, to various restrictions, including restrictions on resale.”

Udell on Gaia

Glue, Gaia, and the services grid: : “Graham Glass, the wizard behind The Mind Electric, is ‘100% sure’ that grid computing is the future. To prepare for it he’s building Gaia, which in its first incarnation will be used to do simple, lightweight clustering and load-balancing of web services. Those services, initially, will be Java-based and written in TME’s SOAP toolkit.”

May be interesting to look at Gaia as we think about how to build a component framework for the enterprise applications using Web Services. We can think of the Thick Servers as forming a P2P network, a computing grid.

TECH TALK: Rethinking Enterprise Software: Whole Solution for USD 20 a month (Part 4)

Let us work out the total computing costs for a 40-person enterprise. One Thick Server can support 40 Thin Clients. I have assumed that the cost of a Thin Client is USD 150 (instead of the USD 100 discussed above). We should look at the costs over a 3-year-periodthe assumption being that even the SMEs will upgrade their systems every 3 years (thus lagging in hardware by no more than 3 years). So, the total costs taken over a 3-year period are:

  • Thin Client: USD 150 one-time, with USD 15 for support per annum. Also, assume there is 1 spare TC kept for every 10 TCs. Thus, the total cost over 36 months is USD 200 per user, or about USD 6 per user per month (pupm).
  • Thick Server: A standard, new desktop can be used as a Thick Server, with the additional need for RAID (to mirror the data on the hard disks). Add to this support for backup. The Thick Server will cost about USD 1,200 with an additional 10% per annum for support. Thus, the total pupm cost becomes USD 1 (or a total of USD 1,440 over 36 months and 40 users).
  • Internet Connectivity: Assuming USD 40 per month for Internet access gives a pupm cost of USD 1.
  • Networking: Putting in place the LAN with the hubs/switches will cost the equivalent of USD 1 pupm.
  • Training and Support: USD 1 pupm.
  • Software: The base modules discussed earlier Enterprise Core and OS, with the Digital Dashboard and Visual Biz-ic should cost no more USD 5 pupm. (If enterprises want specialized components, then they would need to pay extra. This also creates the incentive for software developers to build on the base platform.)

Thus, the aggregate costs for an installation of 40 computers come to USD 6+1+1+1+1+5 = USD 15 pupm. Add to this the costs of marketing and margins for the channels and it should be economically viable to build a profitable business by selling the whole solution for USD 20 pupm. The economic model may be quite simplistic, but one which I think is definitely achievable. Thus, as a whole, the 40-person enterprise pays USD 9,600 (or Rs 5 lakhs) per annum for its entire computing and communications needs. Any company earning a profit of USD 50,000 (or Rs 25 lakhs) can definitely afford to invest 20% of its profits on technology to become a real-time enterprise.

Whats new here is not the technology, but the business model. It has taken the concept of a utility to its extreme, creating the ultimate SME Tech Utility with hardware, software and support being offered as services. Therein lies the trick. For example, if we were to sell the Thin Client for USD 150, there would be few buyers who wants to buy a second-hand computer? But by offering computing as a service (and the TC on a subscription basis as part of a larger solution), IT managers will no longer look under the hood. The analogy is with cars. When we buy cars, we want to make sure they are perfect. When we rent cars (or pay for the service of transportation), we are no longer so concerned as long as the car works and the rental company offers to cover all problems, if any. For SMEs, technology needs to become a utility this will lead to mass-scale adoption by a whole new generation of users.

This then is the overall vision for making technology a utility for SMEs in emerging markets. At its heart is enterprise software built out of inter-changeable components and to standards by a distributed network of factories in the form of the independent software vendors. The software itself is only part of the solution. What SMEs need is a whole solution. All the elements of the solution are actually available today. No new technology needs to be developed to deploy them. What is needed is innovative thinking in ensuring that the whole is truly greater than the sum of the parts.

As we think about how to create new solutions, it would be good to remember these words from Esther Dyson:

In the world we live in, mathematicians and investors have become ever better at calculating risks, assessing outcomes, laying out possible scenarios. But real economic progress comes from taking challenges, not risks, and building something fantastic despite the odds, because you know you’re smarter and more dedicated and more persistent, and you can gather and lead a better team, than any rational calculation would indicate. That’s how new businesses get built, new markets get opened, new value gets created.

And real political, social and ethical progress, likewise, comes not just from negotiating a carefully calibrated “win-win” balance-of-power compromise, matching move for move, but from taking the lead, challenging the other guy to follow, showing the way forward. We make progress by stretching the imagination and doing things we won’t regret. When you cannot predict consequences, then you need to consider your conscience and do what’s right.

We need not calculation, but courage!

The future of Software: Anil Gadre

An excellent article by the Solaris boss in on the present future of software:

Computer operating systems are irrelevant, or should be, to most people.

To home and office users alike, it’s the applications that run on top of the operating system that really count. To the developers who create applications and Web services, it’s the middleware–the application server, directory, and so on–that counts.

The only people who should be concerned with the operating system are the chief information officer and folks who manage network resources. They’re the ones who have to deal with questions of availability, security and scalability–the capacity to grow without adding undue complexity.

Developers shouldn’t need to think about the OS; they should be able to aim higher, at a new software category I call the “service-delivery platform.”

The aim of the service-delivery platform is to make it just as easy for developers to create a large-scale service as it was to create a single shrink-wrapped program for a standalone PC. Developers need to know that certain components are always going to be there–a directory, a network file service, an application server. Those are the components of the service-delivery platform.

Although part of the base platform, these components may also come from a variety of companies offering open-standards-based technology, so long as they present a set of core services that developers can count on–and no proprietary extensions or libraries to cause porting problems.

Exactly the way we need to think of the Emergic platform.

An Entrepreneur’s Walk

I was reflecting recently on the past 3 months, since we began working on Emergic. At that time, there were a number of ideas. I had a set of names for the things we wanted to do, but had little clue on how we would go about doing things. [My May 10 posts]

What we’ve done in the past 3 months is started on many parallel initiatives, which are taking us in the general direction of where we want to go. I don’t have a detailed road map of the terrain. So, as I stand, I walk around on many of the roads and see if they are going in the general direction of where we want to go. This is like taking some small initiatives. As one walks, we see either some clearing or more fog. So, one has to either retrace back, or move along further. These are the decisions an entrepreneur has to make many times a day.

It’s what I have called “An Entrepreneur’s Walk”. The only way one will reach the destination (or close to it) is first by deciding to walk (standing still won’t take one anywhere). Each road may seem equally hazy at first, but walk a little and the landscape will change. It is then the entrepreneur has to decide whether to walk further or go back and try a different road. Some roads will meet a little further down and make a bigger road. Some roads will lead to a dead-end. It is only by walking around on multiple roads that the entrepreneur will make progress.

That’s why it is so important to get rid of the fear of the unknown. It is trial-and-error, but what’s always guiding the entrepreneur is that inner intuition, the hidden hand of God, who tests one periodically to see how determined and confident one is of one’s mission. So, an entrepreneur must leave fear aside and not be afraid of making mistakes. Out of these small mistakes will come bigger successes, and slowly, a map of the landscape will start building.

These are the “walks” that an entrepreneur must enjoy.