Linux in the Enterprise: Opinion

“Our competitor is not proprietary software, it is non-consumption”

As you sit in Rajesh Jain’s office in Mumbai, the view is dominated by smokestacks. The mills that these smokestacks belong to are being dismantled and replaced rapidly by new age companies like Rajesh Jain’s Netcore. Too expensive to demolish, the smokestacks themselves remain as witnesses of a nation’s slow transformation from the industrial era to the digital era. From his perch within the Netcore office, Jain dreams of another transformation where IT reaches “every desktop and every household within India.” Jain is a visionary who set up India’s first portal, Indiaworld, at a time when the Internet was unknown and commercial ISPs were a distant mirage. After the spectacular sale of Indiaworld to Satyam that set off the Internet mania in India, Jain shunned the limelight to focus on Netcore, a company that builds ‘thin-client, thick server’ computing solutions targeted at SMEs, education and community information centres in rural areas. A prolific and thought-provoking writer, Jain has been arguing consistently through his blog on Emergic.org that we must bring down the cost of computing and devise innovative methods to take IT to the Indian masses. Venkatesh Hariharan interviews him to find out why Jain argues so passionately for the adoption of Linux in India

Indiaworld was one of the earliest users of Linux. Why Linux?

We started using Linux in 1994. We were hosted on Linux and all our development was done on Linux. At that time I wasn’t much aware of Linux but our technical team decided to use Linux and I went along with what they said.

Then you started Netcore.

Netcore was started in 1998 to look at messaging solutions. We started it to create websites for corporates and received feedback that they need something to manage their mail. So we created a Linux-based messaging server and that’s what we continue to do today among other things.

What are the other activities of Netcore?

Right now, most of our revenues come from our messaging server. What we have recently done is create a Linux based ‘thin client, thick server’ model. We will be launching this more aggressively. We have been using this internally for more than a year now. The idea is to bring down the total cost of computing and improve administration and manageability of the infrastructure. We plan to push this aggressively at small and medium enterprises (SMEs). We can offer SMEs a Linux server with all the applications that they need, right from messaging, desktop computing, accounting, CRM, knowledge management, file server, print server, etc.

Would all these be open-source software?

Most of these are built on open-source applications, a few we have built on our own. The ones we have built on our own are things like the digital dashboard and RSS/Infoaggregator and Event Horizon, which can create RSS streams from ODBC-compliant databases. Our focus has been on integrating all these into a seamless package. For example, we have created a single log-in which would work across multiple packages. The idea is to give the organisation one server that is remotely managed. The applications are on the server and the clients are low-cost, low-configuration clients.

How many users does a server support?

It scales anywhere from five users to 300 users. Today, we can leverage Moore’s Law on the server. There is no limit to how many users can be supported. We can also use multiple servers by clustering.

Most of Netcore’s work has been Linux-based. Any specific reason for that?

When I look at the numbers in India in terms of the hardware offtake and piracy levels, what it boils down to is the following: The choice that organisations face is between non-consumption and piracy. On one side there are expensive solutions available and you will not use those solutions. In some cases, whenever you can pirate, you will pirate and we have high levels of piracy in India.

Our belief is that if you can make things available at an affordable price, there will be a dramatic growth in the usage of computing. It will create a domestic market which we sorely lack. While we have a successful exports business in software, we definitely need to have a domestic market. If Indian industry is to be competitive, we need to be big users of technology. We need to make complete solutions available at an affordable price and with no compromise on the applications front.

Just because we cannot afford things, we don’t want to tell people to compromise on applications and performance. This is where Linux provides a very strong platform—especially Linux which is running on the server so that your desktops can be of much lower cost. Instead of three new desktops, you can now give computers to 10 more people in your company for the same investment.

The penetration of technology in Indian companies is so low—5-10 percent is what we are at. The installed base is just seven to eight million. Basic applications like messaging, etc, will not be effective if only a small number of people use it.

The idea is—what does it take to get a computer on every desktop and accessible to every family in the country? This is where a combination of three things is key. This is what I call the 5K PC ecosystem. This consists of the thin-client, server-centric computing and open source software.

What this does is bring down the input costs for computing. You can’t sell at a low cost if input costs are high. At the same time, we do not want a nation of pirates. People can pirate a few applications but most applications they cannot pirate. Therefore they cannot use them and are doing things inefficiently. So they are caught in a technology trap.

Indians have made few contributions to the open-source code base. Why is that so?

One reason is that open source is not a significant usage platform in engineering colleges today. We still have a small group of people developing on Linux. These are small shops that don’t have the wherewithal to be part of the production system. The two large entities that can be producers are the engineering colleges and the large software companies. They have, at any point in time, 10-20 percent of people on the bench, who instead of being idle can be put on short three-month projects. People then get trained on something that is state-of-the-art and that people actually use and something that goes back to the community. Multiply 10 percent of the Indian software industry that we can leverage for open-source projects and the number of projects is huge. We can, for example, contribute to Wine, LTSP and any number of projects.

We need a few people who will be there for an extended period of time—the project managers, and then we can have people who do the coding and are there for a shorter period of time, say two to three months. As long as there is continuity maintained by project managers we can definitely take care of production.

Today, we are caught in a low-equilibrium solution. There is an equilibrium but it is in a bad state. Today we have students in engineering colleges who are using ‘free’ software, but they are using pirated ‘free’ software, like Windows. They are not using Linux. Part of the problem is also that the education system hardcodes Windows into the system. The education system is something that needs rectification. It should be made platform neutral, OS neutral.

As a nation, we should stand up against piracy. If I am a company which is making manufactured goods, I don’t want anyone making rip-offs of my products. An anti-piracy stance enables private companies to come up and fill the need with low-cost products. Today, companies do not develop products because they know their products will be pirated. The domestic market is small and out of that, they will get no more than 5-10 percent of the customers, so why bother?

If they know that there is a significant user base and make affordable solutions, the market grows. People start paying for products. Automatically you have entrepreneurs coming in. The rest of the ecosystem will start building up in engineering colleges.

What needs to be done?

There are three key markets which can play the facilitation role. SMEs, engineering colleges and rural areas where there is an increasing use of ICTs happening.

For SMEs we need core applications which go beyond just office applications like word processing. Accounting, CRM, sales force automation, etc, a lot of these applications are available on Linux and we are trying to put them together so that there is a full-fledged suite for businesses. Once these people start adopting Linux, engineering colleges need to provide the supply of people who can build on these applications. They can take up projects for different verticals. Colleges in Pune can, for example, build applications for auto companies. Today, we have the building blocks. You have an application server, Jboss; you have a database in Postgres, Web services, business process standards like EBXML, RosettaNet, etc, which could be used. Today, for the first time, you have software standards, you have business process standards and Web services, which can standardise business processes between companies.

Companies need to take the first step and say, we will not buy pirated software. If we can bring down the total cost of ownership over a 36-month period to something like Rs 600 a month, that’s something like Rs 21,000 over three years for hardware, software, support and training. That means anyone earning Rs 6,000 can be made 10 percent more productive and the cost of the system is paid back almost immediately.

But thin-client systems have been spoken about for more than six years now?

Part of the reason has been that the focus so far has been not as much on cost reduction as on reducing the administration cost. What we are saying is that by itself, thin client will not reduce cost. Citrix-based solutions will not reduce cost because if they continue to use Windows, the savings that happen on hardware are taken up by the Citrix license fee costs and you still need the Microsoft licenses for Windows and Office. What is required is for hardware costs and software costs to come down simultaneously. Along with that, we need the remote manageability of servers and desktops. And, who will push these things? Will it be the channels or the resellers? And it will not be easy to get them trained on Linux. So if you can simplify an enterprise’s architecture by running all the applications on the server, and the server can be very easily managed remotely, SMEs can easily adopt these solutions.

Lastly, where do you see Linux playing a role in the future of India?

We need to get up to 10 million PCs a year. The PC is a disruptive innovation and we are not using enough of it. At 10 million PCs a year, we may end up paying $300-400 million a year to Microsoft and that’s not practical. We have to look at those who are not consumers today. The next 50 million users will come from there and very little legacy exists there. Our competition is not proprietary software, it is non-consumption. We need to take what Bill Gates said in the American context—“A computer on every desk and in every pocket”—and translate that into an Indian context—“A computer on every desk and accessible to every Indian family.” We cannot afford dollar-denominated technology, so we should leverage Moore’s law on the server and leverage our strengths in the software industry through open source.

Linux in the Enterprise: Opinion

“Our competitor is not proprietary software, it is non-consumption”

As you sit in Rajesh Jain’s office in Mumbai, the view is dominated by smokestacks. The mills that these smokestacks belong to are being dismantled and replaced rapidly by new age companies like Rajesh Jain’s Netcore. Too expensive to demolish, the smokestacks themselves remain as witnesses of a nation’s slow transformation from the industrial era to the digital era. From his perch within the Netcore office, Jain dreams of another transformation where IT reaches “every desktop and every household within India.” Jain is a visionary who set up India’s first portal, Indiaworld, at a time when the Internet was unknown and commercial ISPs were a distant mirage. After the spectacular sale of Indiaworld to Satyam that set off the Internet mania in India, Jain shunned the limelight to focus on Netcore, a company that builds ‘thin-client, thick server’ computing solutions targeted at SMEs, education and community information centres in rural areas. A prolific and thought-provoking writer, Jain has been arguing consistently through his blog on Emergic.org that we must bring down the cost of computing and devise innovative methods to take IT to the Indian masses. Venkatesh Hariharan interviews him to find out why Jain argues so passionately for the adoption of Linux in India

Indiaworld was one of the earliest users of Linux. Why Linux?

We started using Linux in 1994. We were hosted on Linux and all our development was done on Linux. At that time I wasn’t much aware of Linux but our technical team decided to use Linux and I went along with what they said.

Then you started Netcore.

Netcore was started in 1998 to look at messaging solutions. We started it to create websites for corporates and received feedback that they need something to manage their mail. So we created a Linux-based messaging server and that’s what we continue to do today among other things.

What are the other activities of Netcore?

Right now, most of our revenues come from our messaging server. What we have recently done is create a Linux based ‘thin client, thick server’ model. We will be launching this more aggressively. We have been using this internally for more than a year now. The idea is to bring down the total cost of computing and improve administration and manageability of the infrastructure. We plan to push this aggressively at small and medium enterprises (SMEs). We can offer SMEs a Linux server with all the applications that they need, right from messaging, desktop computing, accounting, CRM, knowledge management, file server, print server, etc.

Would all these be open-source software?

Most of these are built on open-source applications, a few we have built on our own. The ones we have built on our own are things like the digital dashboard and RSS/Infoaggregator and Event Horizon, which can create RSS streams from ODBC-compliant databases. Our focus has been on integrating all these into a seamless package. For example, we have created a single log-in which would work across multiple packages. The idea is to give the organisation one server that is remotely managed. The applications are on the server and the clients are low-cost, low-configuration clients.

How many users does a server support?

It scales anywhere from five users to 300 users. Today, we can leverage Moore’s Law on the server. There is no limit to how many users can be supported. We can also use multiple servers by clustering.

Most of Netcore’s work has been Linux-based. Any specific reason for that?

When I look at the numbers in India in terms of the hardware offtake and piracy levels, what it boils down to is the following: The choice that organisations face is between non-consumption and piracy. On one side there are expensive solutions available and you will not use those solutions. In some cases, whenever you can pirate, you will pirate and we have high levels of piracy in India.

Our belief is that if you can make things available at an affordable price, there will be a dramatic growth in the usage of computing. It will create a domestic market which we sorely lack. While we have a successful exports business in software, we definitely need to have a domestic market. If Indian industry is to be competitive, we need to be big users of technology. We need to make complete solutions available at an affordable price and with no compromise on the applications front.

Just because we cannot afford things, we don’t want to tell people to compromise on applications and performance. This is where Linux provides a very strong platform—especially Linux which is running on the server so that your desktops can be of much lower cost. Instead of three new desktops, you can now give computers to 10 more people in your company for the same investment.

The penetration of technology in Indian companies is so low—5-10 percent is what we are at. The installed base is just seven to eight million. Basic applications like messaging, etc, will not be effective if only a small number of people use it.

The idea is—what does it take to get a computer on every desktop and accessible to every family in the country? This is where a combination of three things is key. This is what I call the 5K PC ecosystem. This consists of the thin-client, server-centric computing and open source software.

What this does is bring down the input costs for computing. You can’t sell at a low cost if input costs are high. At the same time, we do not want a nation of pirates. People can pirate a few applications but most applications they cannot pirate. Therefore they cannot use them and are doing things inefficiently. So they are caught in a technology trap.

Indians have made few contributions to the open-source code base. Why is that so?

One reason is that open source is not a significant usage platform in engineering colleges today. We still have a small group of people developing on Linux. These are small shops that don’t have the wherewithal to be part of the production system. The two large entities that can be producers are the engineering colleges and the large software companies. They have, at any point in time, 10-20 percent of people on the bench, who instead of being idle can be put on short three-month projects. People then get trained on something that is state-of-the-art and that people actually use and something that goes back to the community. Multiply 10 percent of the Indian software industry that we can leverage for open-source projects and the number of projects is huge. We can, for example, contribute to Wine, LTSP and any number of projects.

We need a few people who will be there for an extended period of time—the project managers, and then we can have people who do the coding and are there for a shorter period of time, say two to three months. As long as there is continuity maintained by project managers we can definitely take care of production.

Today, we are caught in a low-equilibrium solution. There is an equilibrium but it is in a bad state. Today we have students in engineering colleges who are using ‘free’ software, but they are using pirated ‘free’ software, like Windows. They are not using Linux. Part of the problem is also that the education system hardcodes Windows into the system. The education system is something that needs rectification. It should be made platform neutral, OS neutral.

As a nation, we should stand up against piracy. If I am a company which is making manufactured goods, I don’t want anyone making rip-offs of my products. An anti-piracy stance enables private companies to come up and fill the need with low-cost products. Today, companies do not develop products because they know their products will be pirated. The domestic market is small and out of that, they will get no more than 5-10 percent of the customers, so why bother?

If they know that there is a significant user base and make affordable solutions, the market grows. People start paying for products. Automatically you have entrepreneurs coming in. The rest of the ecosystem will start building up in engineering colleges.

What needs to be done?

There are three key markets which can play the facilitation role. SMEs, engineering colleges and rural areas where there is an increasing use of ICTs happening.

For SMEs we need core applications which go beyond just office applications like word processing. Accounting, CRM, sales force automation, etc, a lot of these applications are available on Linux and we are trying to put them together so that there is a full-fledged suite for businesses. Once these people start adopting Linux, engineering colleges need to provide the supply of people who can build on these applications. They can take up projects for different verticals. Colleges in Pune can, for example, build applications for auto companies. Today, we have the building blocks. You have an application server, Jboss; you have a database in Postgres, Web services, business process standards like EBXML, RosettaNet, etc, which could be used. Today, for the first time, you have software standards, you have business process standards and Web services, which can standardise business processes between companies.

Companies need to take the first step and say, we will not buy pirated software. If we can bring down the total cost of ownership over a 36-month period to something like Rs 600 a month, that’s something like Rs 21,000 over three years for hardware, software, support and training. That means anyone earning Rs 6,000 can be made 10 percent more productive and the cost of the system is paid back almost immediately.

But thin-client systems have been spoken about for more than six years now?

Part of the reason has been that the focus so far has been not as much on cost reduction as on reducing the administration cost. What we are saying is that by itself, thin client will not reduce cost. Citrix-based solutions will not reduce cost because if they continue to use Windows, the savings that happen on hardware are taken up by the Citrix license fee costs and you still need the Microsoft licenses for Windows and Office. What is required is for hardware costs and software costs to come down simultaneously. Along with that, we need the remote manageability of servers and desktops. And, who will push these things? Will it be the channels or the resellers? And it will not be easy to get them trained on Linux. So if you can simplify an enterprise’s architecture by running all the applications on the server, and the server can be very easily managed remotely, SMEs can easily adopt these solutions.

Lastly, where do you see Linux playing a role in the future of India?

We need to get up to 10 million PCs a year. The PC is a disruptive innovation and we are not using enough of it. At 10 million PCs a year, we may end up paying $300-400 million a year to Microsoft and that’s not practical. We have to look at those who are not consumers today. The next 50 million users will come from there and very little legacy exists there. Our competition is not proprietary software, it is non-consumption. We need to take what Bill Gates said in the American context—“A computer on every desk and in every pocket”—and translate that into an Indian context—“A computer on every desk and accessible to every Indian family.” We cannot afford dollar-denominated technology, so we should leverage Moore’s law on the server and leverage our strengths in the software industry through open source.