An Affordable Alternative

At least one of the top-selling banking software solutions works only on the MS-Windows platform. This means that banks considering the solution have to necessarily use an infrastructure that comprises thick, new desktops and MS-Windows as the operating system. Built into this assumption is the need for its customers to upgrade desktops every three to four years. Are there alternatives?

One of the large banks in India recently tendered for a messaging solution. The tender was open only for IBM (Lotus Notes), Microsoft (Exchange) and Novell (Groupwise). Are there alternatives?

A large financial services organization is considering upgrades of its desktops - all running the MS-Windows and MS-Office organization. The total cost of ownership over a four-year period per user will work out to Rs 100,000 ($2,000), by no means a pittance. Are there alternatives?

India's banking, financial services and insurance (BFSI) industry is the new cash guzzler. Technology is rapidly becoming the oil of the 21st century for India. Investments in hardware and software are being made in large quantities as organizations modernize and want to use IT as both the foundation for their business and for competitive differentiation. But in doing so, we do not realize the cost at which this is happening.

Given infinite or large resources, there is no question that organizations in India would want to use the best technology available in the world. Price is often used as an indicator to grade a technology solution. So, the desire by Indian banking institutions to spend on installing a world-class IT infrastructure is understandable.

Unfortunately, this is not an ideal world and resources are not infinite. Indian banks earn in rupees and spend in dollars for technology. Most technology happens to be dollar-denominated, with the result that the Indian price points can appear quite expensive. Organizations are then faced with two options: either to go ahead without the corresponding dollar-denominated earning capability, or to limit technology investments.

If it was just a single bank, then it might not matter because customers would not have a choice. But there are multiple banks competing for business. So, as one starts spending on the best dollar-denominated technology, so does the other, and then the next, and so on. Who wins? The technology vendors. Who loses? The non-spenders. The IT-spending banks end up neutralizing each other's technology impact, and customers get lower returns on their monies in the bank since a portion of it is being used in an "IT arms race". 

No bank would like to compromise on the technology architecture that it builds, but are there alternatives that can create a comparable technology base, but at more affordable price points?






The solution does not lie in not adopting technology. In no way do we want Indian banks to be behind or technologically inferior to their global counterparts. So, what then is the alternative?

The Indian BFSI players are of two types: the ones at the "top of the pyramid" that are big enough to be able to easily afford the dollar-denominated tech spends, and the ones who are not able to but have to spend to match up to competition. The alternatives are needed more for the second category than the first.

No single bank has the ability to put together an alternative platform, but if some of them cooperate, then it is doable. The challenge, therefore, is to put together an "affordable technology infrastructure commons" (ATIC) on which different banks can build their services. This is akin to the role that an operating system plays in a computer: it encapsulates all the commonly used elements into a standardized interface called the applications programming interface or API so as to make it easier and cheaper for others to build their applications on top.

The difference is that, in this case, the ATIC provides a hardware and software abstraction layer, which brings down the cost of computing significantly, without compromising on functionality. So, what does the ATIC comprise? There are three building blocks for the ATIC: thin clients, server-centric computing and open-source software.

Thin clients are PCs without local storage that undertake very limited local processing jobs. They are, in essence, network computers or PC terminals. They "light up" in the presence of the network and are useless without it. These thin clients can either be the older, existing computers or newer, low-cost and low-configuration machines. Thin clients make up the "Rs 5,000 PC (5KPC) ecosystem". How does one get PCs for Rs 5,000?

New thin clients (PCs without disks, lower memory and processor configurations, and inclusive of a monitor) are available for about Rs 10,000. In fact, the base configuration needed is no more than 100 Mhz in processing power and 32 MB RAM. The machines will boot off a server on the network. So, you must be thinking, how does that get us to the 5KPC? Simple. Look around the bank.

Many banks have existing desktops that could be three to five years old. Ordinarily, these would need to be upgraded. The resale value of these desktops is extremely limited. These desktops are perfectly functional systems - the problem is that over time, the need for faster computers is being driven by newer software, which needs greater resources. So, banks are left with little option but to upgrade. Imagine now, if the same desktops could be used without any changes or upgrades to provide the performance of a new computer. The new incremental investment on these desktops would be zero.

So, we have Rs 10,000 for new thin clients and Rs 0 for existing desktops. Assuming a roughly even split between new purchases and older machines, we get an average cost of Rs 5,000. Welcome to the 5KPC world.

The 5KPC never needs to be upgraded - all the processing happens on the server in the network. This is one of the reasons its administration is simplified. Think of the telephone - there is rarely a need to call customer support. In the case of the 5KPC, network connectivity is what makes it come alive - the network provides the digital dial tone. This simplification also ensures that the 5KPC is a zero-maintenance device. This is especially useful when these computers are deployed in remote locations (bank branches are everywhere). The 5KPC is, thus, a "sealed endpoint" - if it does not switch on, it needs to be replaced - there is no debugging it.

Thin clients need thick servers to do the processing and storage. The "thick server" that we refer to here can be of two types: it can be a single, new desktop computer with enhanced memory and two hard disks with real-time mirroring of data ("software RAID"), or a collection of clustered desktop machines. Think of these as inexpensive "blade servers" with network-attached storage. The second solution circumvents the single point of failure problem inherent in the first option, thus offering greater scalability and reliability. The investment on the server would be about Rs 1,500 to Rs 3,000 per client attached to the system.

The third element of the solution is the software. The base for the client and the server is Linux and other open-source applications. The basic set of applications on the desktop include an email client (Ximian's Evolution), a desktop productivity suite (OpenOffice, which can read and write files in DOC, XLS and PPT file formats), a web browser (Mozilla or its lightweight variants), an instant messaging client (GAIM) that provides interoperability with existing IM clients (AOL, ICQ, MSN and Yahoo), and a PDF reader (Adobe's Acrobat). All these applications are available for free on Linux.

Applications run on the server and are displayed on the 5KPC using either a terminal-server application like LTSP (Linux terminal server project, which runs an X server on the client) or vnc (virtual network computer). vnc, created by AT&T Labs, is "a remote display system that allows you to view a computing 'desktop' environment not only on the machine where it is running, but from anywhere on the Internet and from a wide variety of machine architectures".

The idea of carrying out processing on the server and sending keystrokes and mouse clicks from the user and getting the updated screen from the server is not a new idea: running applications on the server over low-speed connections is already being done - Citrix has a solution that works in the Windows world. What is new here is using a Linux desktop to not only cut desktop and server hardware costs, but also software costs.

The big opportunity for the ATIC is at the branch-level. Each branch can have a 5KPC for every employee, connected to a thick server. The users now get the performance of a new "thick desktop", the look-and-feel of a Windows-like interface, the full complement of applications (email, browser, IM, office suite) without the attendant problems of having to upgrade every few years. In addition, support is simplified dramatically because the client computers do not need any support and the thick servers at the branches can be managed centrally.

What ATIC does is bring down the single biggest impediment to computerization: the high cost of hardware and proprietary software (MS-Windows and MS-Office). By using the ATIC architecture, the estimated cost savings per user will be Rs 40,000. Multiply these savings by a few thousand users, add to the amount the other benefits of lower administrative costs, lesser virus worries and simpler application upgrades, and the benefits of an ATIC architecture become apparent.

Open-source software has much wider usage. So far, we are familiar with the use of Linux for mail (Sendmail), proxy (Squid), file and print (Samba) and web (Apache) servers. There are lots more. There are databases like PostgreSQL and MySQL, there is an application server JBoss, there are open-source CRM and sales management applications like OpenCRM and Relata. In fact, there are open-source applications available for almost everything that one can think of. The problem is that many of these applications may not be "finished" or may have inadequate documentation and support. This is the opportunity for India's software industry to build on top of and around these applications.

The one issue that has not been addressed so far is that of core banking applications. Most applications today are Windows-based or need Internet Explorer. This is where cooperating banks need to put pressure on vendors to make applications work on Linux or other browsers. Many database vendors and applications work on Linux (or another Unix variant) at the server-level. In fact, even if banks finance the cost of getting applications re-written by vendors to support Linux and other browsers, it will be well worth the investment.

I am not suggesting that Indian banks do away with proprietary applications or slash IT budgets. No. What they should look at are hybrid solutions - a mix of open-source and proprietary applications will create a win-win solution for everyone.

The most important requirement to make this happen is an open mind on the part of the technology decision-makers in banks.

In fact, one of the side effects of this could very well be that proprietary solution vendors start offering their products are at more reasonable prices rather than dollar-denominated or monopolistic pricing. Banks should use their IT budgets to deepen the penetration of technology both in terms of employees having access to computing as well as the type of applications in use (for example, investments in knowledge management and data mining software).

These ideas may seem quite radical and counter-intuitive. After all, the banking and the financial services sectors make plenty of money. Why should they worry about reducing costs by going in for technology that may have not been used elsewhere in the world? Why should they be the guinea pigs? In fact, from the point of view of most managers, this is just the solution they should not want to try out - it is different, it is not Windows, it may be risky, and most importantly, the money being spent is not their money.

The issues we need to consider as a nation are: optimum utilization of limited resources and self-reliance. Expenditures on technology may not seem much when we as a nation are buying 2 million computers a year and running pirated software on more than 70% of these. Our nation of a billion people needs a technology infrastructure as much as it needs roads, power, food and water, if we are to realize the dream of being a developed nation in the coming years. The need of the hour is disruptive innovations and not status quo.

Imagine the situation a few years from now. If demand in India rises to 10 million computers a year, then just our software spend on Microsoft's Windows and Office will be $4 billion (Rs 20,000 crore) per annum. (Microsoft's profit margins run at 80% to 85%. Little wonder then that Microsoft is prepared to "gift" us $400 million now to hardcode its Windows and Office into our schools.) Add to that the spend on new hardware when we could have done with older, refurbished desktops, and our technology import bill is likely to exceed $8 billion (Rs 40,000

crore).

Considering that the Indian BFSI industry is one of the largest tech spenders, it needs to show the way in adopting alternative technologies. The near term is not going to be easy. But this is the one industry that has the clout to make a difference and create a ripple effect across the rest of India.

The need of the hour is for Indian banks, especially the smaller and medium-sized banks, to cooperate on adopting ATIC as the base platform. This will help conserve significant amounts of capital that can be spent in branch expansions and providing additional services. Moreover, with a lower but deeper technology spend, they will be able to compete better with the bigger players.

In fact, the mantra for these banks should be a computer on every desktop for every employee. Basic applications like email and instant messaging do not work if half the organization does not have the infrastructure.

As we look ahead, it would be useful to keep President Abdul Kalam's words in mind: "The most unfortunate thing is that

India still seems to believe in proprietary solutions. The further spread of IT, which is influencing the daily life of individuals, would have a devastating effect on society due to any small shift in business practices involving these proprietary solutions. It is precisely for these reasons that open-source software needs to be built which would be cost effective for the entire society. In India, open-source code software will have to come and stay in a big way for the benefit of our billion people."

The author, Rajesh Jain, is managing director of

Netcore Solutions. His weblog is at http://www.emergic.org