Sun's announcement on Linux desktops was disappointing. What Sun basically said is that they will provide a desktop hardware and software solution which will cost USD 300,000 in quantities of 100 to buy and administer over a 5-year-period, compared to USD 1 million for Windows. It will bundle open source software on the desktop.

From a Reuters release (via NYT):

The new desktop, dubbed Mad Hatter at Sun, will probably be a small-sized machine with an average microprocessor, hard disk drive and RAM memory costing about $1,000, including monitor.

The machine will also have a reader for credit-card sized identity cards to ensure security using Java technology and likely will not include a floppy disk drive. Sun officials predicted it would go for sale in the first quarter of 2003.

Sun plans to provide only one version of the computer and sell it in lots of 100 along with a powerful server computer and a services fee. Users could take advantage of free e-mail and word processing software and might use the browser as the primary tool.

Equipment and maintenance costs would be about $300,000 for 100 users over five years, or about $50 per user per month, Sun's software chief, Jonathan Schwartz said, estimating traditional PCs would cost about $1 million, including Microsoft Office software and support over the same period.

More from an AP release (via NYT):

Sun's Linux systems will be sold in increments of 100 units; the company will target customers in industries, such as call centers, retail banks, payment centers, schools and government institutions. Pricing has not been announced. The company expects to sell the computers and levy a monthly subscription fee for service and support.

The desktop systems, set to come to market next year, will employ "military-grade security,'' the company said.

The computers are not meant to compete directly with Dell's Linux-based PCs, which target both individuals and corporations.

"The target demographic is all the people in a fixed-function environment ... who probably do not need a $3,000 desktop,'' Schwartz said.

The desktops will include several free open-source programs: the Linux operating system, a Mozilla Web browser, the GNOME desktop graphical environment and the OpenOffice suite of word processing and database applications.

The per unit cost for five years comes to USD 3,000 for Sun (vs their claim of USD 10,000 for Microsoft). At the proposed price of USD 1,000 for the desktop, it is way too high. And so is the USD 400 per annum for support and software and other stuff.

I think this is 10x more expensive than it should be. Of course, Sun is assuming "total cost of ownership" which means it includes support costs and software upgrades and perhaps more. But the costs are still way too expensive.

Sun had an opportunity to redefine computing and that is where it failed.

Here is what Sun should have said: "We are using our strengths on the server to provide a server-centric computing environment wherein you never need to upgrade your desktops, and in fact, can recycle desktops. You can buy older machines for less than USD 150. We'll provide all the software that you need based on Linux for USD 50 per annum. Assuming a loaded cost on the server for USD 50 per client (one-time), the year 1 hardware infrastructure costs USD 200 per client and the software costs USD 50. Now, over 5 years, assuming we need to replace 1 desktop in every two (it just stops working or whatever else bad happens), our projected 5-year cost per desktop is USD 200+100+250=USD 550. Okay, lets add USD 200 so we make some money and so does everyone else in the chain. So, for 100 Linux desktops (and the servers necessary for them), the 5-year operating cost is USD 75,000."

[All of this is possible: we intend to do just that -- see my earlier postings today on TC-TS Positioning and TC-TS as Tech's Pain Reliever....Just scroll down.]

Instead, Sun has come up with a pure desktop business based on Intel CPUs with open source software. They have no competitive advantage in either of the two areas. Instead, their strength, which they should have played to, is on the server side.

Sun's Thin Client Solution (SunRay) costs USD 500. That is way too much. The way to beat Microsoft is to drop costs of hardware and software -- the price points for users have to be 10-20% of what they are at today. The developed markets may still take time to switch because of legacy and the applications base, but the emerging markets will jump at this. That's the opportunity Sun missed: to create a computing platform for the next 500 million users.

It is almost as if Sun expects Microsoft to be not listening when they made the announcement. What happens if Microsoft makes Windows free tomorrow? That is why one needs the hardware-software as a combo to drop in prices. The only way the hardware prices can drop is to recycle PCs and eliminate the need for upgrades. That is what server-centric computing can do. That's the opportunity Sun missed.

I like Sun as a company. I used Sun workstations when I worked at NYNEX in the late 1980s. A lot of innovations have come out of Sun, and perhaps more will. Sun, like many others, needs to think beyond the US, beyond today's desktops, beyond today's computing models, beyond today's users. That is what they did not do. That's the opportunity that Sun missed.

Following on my previous post on our learnings from the market, here is what we think are Tech's Pain Points ("tyrannies") and how our Thin Client-Thick Server (TC-TS) solution can provide relief:

1. Cost of Proprietary Software: After the hardware, the biggest universal cost is that of Microsoft Windows and Office. Till now, there really has been no alternative. StarOffice has been tried my various companies but one common issue has been that of performance ("too slow"). What TC-TS does is eliminates most of the USD 500 (Rs 25,000) per user cost of Windows and Office because of its use of open source (Linux, KDE, Evolution, Mozilla, OpenOffice, GAIM). OpenOffice is more than good enough for most users. For 30 users, it means a saving of USD 15,000 (Rs 7.5 lakhs). Because it runs OpenOffice on the server, perfomance and the desktop hardware becomes a non-issue (also see point 3 below).

2. Virus Menace: On the Windows desktop, one either needs anti-virus software (which costs additional money), or has to live in constant fear of a virus attack. TC-TS provides a virus-free environment.

3. Hardware Upgrades: Every new version of Windows requires desktops to be upgraded. More money to be spent to run the same applications! By moving computing to the server as TC-TS does, desktops never need to be upgraded. Instead the money saved can be used to provide a new set of users with computers. Now, all the old computers can be used rather than just be disposed off.

4. New Computer Cost: Every new computer costs at least USD 500-600 (Rs 25-30,000). By contrast, a TC (Thin Client) can be an old PC - like a Pentium II, with 233 Mhz processor and 16/32 MB RAM. No hard disk or CD-ROM drive is required. These desktops are available for less than USD 150 (Rs 7,500). Even newer stripped down machins will cost no more than USD 300 (Rs 15,000).

5. Computing for Everyone: This is now no longer a dream. Even if the TC costs are high, a shared set of terminals (TCs) can be set up with privacy guaranteed for each user.

6. Software upgrades: Every time a new version of software is released (or a new internal application needs to be deployed), all desktops need to be upgraded. In the case of TC-TS, the upgrade needs to be applied only on the server.

7. Remote Management: Desktops can be managed and configured from the server, dramatically easing the life of the IT team. Even the desktops of remote branch offices can be managed from the head office (through the Internet and the TS at the branch).

8. Centralised Backups: A big worry with desktops is backing up the individual desktops. With TC-TS, since storage is centralised, backups now become very easy.

9. Windows Applications support: This has been the biggest challenge. We now have solutions for this. For an additonal cost of USD 100 (Rs 5,000), a TC will be able to run almost every Windows application. This protects the legacy software written.

10. Double the Desktop: One issue with the desktop (which I have written about in my Tech Talk series this week on "Rethinking the Desktop") is that we have moved from a stand-alone world to a networked and collaboration-centric world, but our desktop has not changed. The TC-TS Solution offers two desktops, separated only by a single click, thus effectively doubling the usable screen space at no additional cost. (I describe how the two Spaces can be used in my Tech Talk articles).

Overall, the cost savings are significant. Let us look at three scenarios: a Windows desktop, a Linux-only Thin Client, and a Linux Thin Client with full Windows applications support.

Windows Desktop - total cost: USD 900
- Hardware: USD 500
- Windows: USD 100
- Office: USD 300

Linux Thin Client Desktop - total cost: USD 300
- Hardware: USD 150
- Server Load: USD 50 (assuming USD 1500 server for 30 clients)
- TC Software: USD 100 (includes OpenOffice)

Linux Thin Client Desktop with Full Windows support - total cost: 500
- Hardware: USD 150
- Server Load: USD 50 (assuming USD 1500 server for 30 clients)
- TC Software: USD 100 (includes OpenOffice)
- Windows: USD 100
- Windows Emulator: USD 100

Thus, the savings are USD 400-600 (Rs 20,000-Rs 30,000) per desktop. Multiply this by the number of users in the organisation, and the savings can get quite big very quickly. That is the real opportunity for TC-TS, and what we need to highlight.

If the latest hardware exists, on the software front alone, the savings range from USD 50-250 (Rs 2,500-Rs 12,500), excluding hardware upgrades which new Windows versions may need.

Add to all this the other administration and management benefits, and the TC-TS solution now can look like a very powerful desktop alternative. Also, by dropping the cost of computing, it now becomes possible ti imagine new segments where we can deploy computers (eg. hotel rooms). This is the message we need to get out to prospective customers.

TC-TS is both a Pain Reliever and an Opportunities Creator.

For the past few days, I've been meeting with some of our MailServ customers pitching the Thin Client-Thick Server solution to them. The meetings have been disappointing from us getting them to trial TC-TS. This led to some deep thinking on what we were doing wrong in how we were positioning the solution.

There were 3 key things we were doing wrong:
- positioning it more as a technology solution ("look: thin client, thick server; look: server-centric computing; look: Linux desktop") rather than understanding what their problems were and how we could solve them
- positioning as a pure software solution ("we'll give the TC-TS software and you need to get a server with these specifications - dual CPU, 2 hard disks, lots of memory")
- ignoring the dominance of Windows applications in the enterprise (Windows legacy is deep and companies have created lots of custom applications for their use). So, we'd become defensive ("there must be some people who only need the base set of apps...can you think")

This approach was not getting us anywhere. A few face-to-face meetings with prospective clients convinced me of two things:
- TC-TS is a great solution. I believe that even more now!
- Positioning makes a world of difference. We were not putting ourselves in the customer's shoes. We were too enamoured with the technology itself.

So, then as we brainstormed, a few ideas came up:
- we need to position this as an Alternative to the Microsoft Windows-Office desktop with major cost savings
- we need to be able to support Windows
- we need to include the server in the solution (we are really a desktop alternative; how we do the desktop is our problem; TC without TS is irrelevant; customers want a single point of contact)
- we need to emphasise upfront the cost savings and let the customers work out how this fits in. For 30 users, if we can save USD 15,000 (at USD 500 or Rs 25,000 per user), they will come back to us saying - "you know, this is what we want to do with it" (we did find some of the prospects trying to do this - money is obviously very important).

So, based on the above, I've now worked out (a) the Pain Points / Hassles that companies (and especially CIOs) have, and (b) how our TC-TS Solution can help relieve those pains. In other words, think of TC-TS as Tech's Pain Reliever. More on this shortly.

One Screen, Two Spaces. Why?

In the past few years, the Internet has dramatically increased the connections that we are able to make with people, information and applications. For starters, most of us have Hotmail or Yahoo (or another equivalent) email addresses separate from the one we get at work. Concomitant with the email address now comes the presence through an Instant Messaging tool. Whether we like it or now, we've already started donning multiple identities.

One way to separate the identities is to use distinct computers - do one's corporate activities at the workplace computer, and personal activities from the one at home. This is easier said that one. For many, it is not possible to have separate computers. Also, at times, it is not possible to cleanly separate activities as email and IM have extended working hours, and friends and family connect to us during business hours.

So, even as the number of relationships and interactions we maintain has gone up manifold, the screen area available to co-ordinate these has barely increased. In fact, I remember as a software developer in the early 1990s, using two monitors on the desktop - one for writing the programs, and the other for running the application and doing the debugging. Many financial workstations have dual monitors. So, considering the complexity and diversity of information and activities that we need to track, why not think of having two workspaces on the same monitor? At one level, think of the two spaces as doubling the screen area. But more importantly, the two spaces represent the two hats that we wear while using the computer.

The Digital Dashboard comprises of two Spaces - a personal space, which we can designate as "MySpace" and a business-related space, which we can designate as "WorkSpace". The two spaces have an identical set of tools and applications (described yesterday) which can be combined and arranged together differently. A single click from the bottom of the screen (on the taskbar) should toggle between the two spaces. Each space can provide access to information from the other, but what each space does is bring into sharper focus that information and those activities relevant to the specific space.

(On a related note, the idea of two Spaces came up as I was looking at the Linux Thin Client desktop and thinking about what is it that we can leverage from the desktop which is not there on Windows, and which can genuinely give a benefit to users. So, while not having Windows and Office may "disappoint" some users, the second desktop can be one which can "delight" and perhaps, even encourage users to switch from Windows to Linux. Of course, there is a little additional work in terms of switching between two desktop spaces, but I do not think that will be a problem: we seem to handle multiple TV remotes and multiple car steerings quite well! The time has come for multiple Spaces. Let's start with two.)

Tomorrow: Spaces (continued)