February 1999

The Little OS That Could

BY ROBERT VAHID HASHEMIAN

I have good news and bad news for Internet telephony and CTI companies. But first let's find out why, then I'll get to the news. Does the name "Torvalds" ring a bell? If not, I won't hold that against you. Who can remember names these days anyway? It happens to be the last name of Linus Torvalds. Yes - the man credited with the invention of Linux. The now-famous OS started its humble life out of a university teacher's office and quickly spread its wings to take the world by storm. (Actually, Linus Torvalds was just tweaking with task-switching for the i386 architecture in 1991 and unwittingly created the kernel for Linux.)

These days, you'd be hard-pressed to find any technical journal that doesn't prominently mention Linux. In fact, putting Linux in a headline has become the "cool" thing to do for many publications. I won't even be surprised if the baby high-tech Teletubbies do a show on Linux (if you don't know what the Teletubbies are, ask a colleague with a toddler).

My first introduction to Linux was in 1994, in Germany. My wife's college in Darmstadt had a computer room with a few Linux machines, so I got an account through her and played with the systems from time to time. I was pretty well versed in UNIX, so it didn't take me long to find my way around this new OS, and soon I became a fan. But as it happens with many old friends, I changed jobs and got immersed in the Microsoft Windows world.

LINUX GAINS STEAM
I forgot all about Linux - until about a year ago - when I started to hear and read about it in different technical publications. While the news of Linux's momentum was being dismissed as hype by many of my colleagues, I took my old Pentium-90 PC and secretly installed a copy of Linux on it, which I had obtained from one of the well-known Linux companies. To my surprise, the installation was smooth as silk. The Linux vendor had done an excellent job creating a user-friendly installation script coupled with good documentation. In no time, I had this machine on the network rubbing elbows with high-powered Pentium-IIs running BackOffice.

In the next few days, I spent my time scratching the surface of this OS, trying to understand why it has endeared itself to so many in such a short time. Here is my list of things that I found attractive about Linux:

• It is inexpensive. Actually, Linux itself is free and considering its lean structure, it can run on just about any cheap hardware, drastically reducing the overall cost of a system.
• If you like commercial products, Linux fits that bill too. Companies such as Red Hat and Caldera offer commercial versions of Linux with multiple levels of support and maintenance. Of course, you will always get plenty of free support from the thousands of dedicated Linux users and developers online - hungry to answer questions and troubleshoot problems.
• It is stable. While I must reboot Windows every few days, Linux can run for months without any reboots.
• It is faster. Linux is not bogged down by myriad resource-hogging modules. It has a small footprint, leaving plenty of resources for other applications to consume. This is evident by how fast Linux comes online after a reboot (that is, if it needed one).
• It has a robust command line interface. Please don't give me the old argument that GUI is easier. Anyone who has handled Windows NT knows that GUI doesn't help one become a better system manager - you need just as much training and experience in NT as any other OS. I find myself using NT's command window more and more to carry out many tasks, although NT's command line interface is still woefully inadequate.
• The subsystems are truly independent. What are called services in Windows NT are called daemons in Linux. Daemons can be stopped and started without damaging the operating system or other subsystems. While this is claimed to be true of Windows NT services, I have often had to reboot NT because stopping or rebooting a certain service damaged the running OS or other services.
• Its kernel and subsystems can be customized and redeployed. While such modifications require strong programming and system knowledge, they are not outside of the realm of possibility.
• Finally, and perhaps most importantly for corporate use, Linux has garnered major support from such heavyweights as IBM, Intel, Oracle, Compaq, HP, Sun, and even (gasp) Microsoft.

WHO'S MISSING THE TRAIN
You can bet that Linux is here to stay and it will continue to broaden its acceptance as more people discover its benefits… Except for Internet telephony and CTI companies! It just recently occurred to me that although Linux is growing in popularity, I don't hear a squeak about it from the huge number of manufacturers, vendors, and service providers operating in the Internet telephony and CTI industries, and I wondered why. Try doing a search for Linux on the Web sites of companies such as Dialogic, Natural MicroSystems, Nortel Networks, VocalTec, and Brooktrout, and you will see what I am talking about.

I assume if one asks about this, the canned answer will be, "We are watching Linux's development and will address it as it becomes necessary." I suspect that after years of being pressured from customers and publications (not unlike this one) to enter the Windows NT market and finally giving in, they are just too exhausted (financially, technically, and mentally) to attack another market. And who am I to blame them? What I hope is for the industry to make serious preparation for supporting Linux. Here is my list of items to prepare:

• Learn Linux. Have at least a few people who are proficient at it.
• Find out what development tools are available and begin experimenting with them.
• Make the software as OS-independent as possible. This will allow a much easier migration in the future.

HOP ABOARD BEFORE IT'S TOO LATE
And now - finally - the good and the bad news. The good news is that:

• Linux is not a threat to Windows NT or other operating systems yet. There is enough time to develop a Linux strategy.
• Systems running on Linux have several advantages:
  1. They are easier to develop and maintain,
  2. They run more efficiently, and
  3. They offer a lower initial investment and TCO (Total Cost of Ownership) to end-users.
• Many Internet telephony and CTI vendors had their start on UNIX systems. This will become an immense benefit as they consider support for Linux. They have already been there and done that.

And the bad news should be evident by now. If you haven't considered Linux yet in your strategy, you could be taking a big gamble in contrast to others who may beat you to it. By some accounts, there are about 10 million Linux users around. Personally, I think the number is higher. Many more Linux fans will be joining the corporate world as they graduate from colleges and universities. Will you be ready to serve them on their turf?

Robert Vahid Hashemian provides us with a healthy dose of reality each week in his Reality Check column. Robert currently holds the position of Webmaster for TMCnet - your online resource for CTI, Internet telephony, and call center solutions. He can be reached at rhashemian@tmcnet.com.