P2P? Not for me

Peer-to-peer computing has been generating a lot of buzz lately. Instant messaging and Napster boosted the visibility of P2P–in which two computers exchange information without an intervening server–even though Napster’s high-profile music-sharing service isn’t exactly serverless. At the same time, Ray Ozzie’s Groove Networks aims to move the model to the corporate world.

Sorry, folks–P2P isn’t going make it in the enterprise. P2P had its chance a dozen years ago and never scaled beyond the workgroup. P2P had its chance a dozen years ago or more and never scaled beyond the workgroup. Anyone remember Artisoft? Ten years ago its LANtastic was the premier peer-to-peer network operating system. That was before Microsoft added P2P to Windows for Workgroups and drove the company completely out of the market. (LANtastic is still limping along under the corporate umbrella of network connectivity vendor SpartaCom Technologies.)

Novell, which uses a client/server model for its NetWare network OS, spanked Artisoft’s butt in the battle for corporate desktops. Why? Because putting your critical resources in one carefully watched basket makes business sense.

Management is one part of why client/server is a better model for corporations. Centrally located resources are easier to manage than assets distributed across a company. Efficient backup, fault tolerance, and security are easier to manage on a centralized server.

Cost is another part. If each client acts as a server for every other client, you need faster desktop hardware and a faster network infrastructure.

P2P makes a fine consumer model, especially given the pervasive nature of the Internet. It’s useful for exchanging files and messages, à la Napster and AOL’s Instant Messenger, neither of which is an especially useful business tool. It might be useful to employ clients for their unused CPU cycles, like SETI@Home, but so far no one has put forth a workable product in that area, and when it does come, it won’t really be peer-to-peer, but “server/client”–that is, a server will coordinate all the clients’ resources.

Today, server-based computing is firmly entrenched in the business computing world. Even workgroup collaboration platforms, which might seem a natural for P2P applications, generally sport a client/server architecture. If there are going to be any pervasive changes to that model, it will be from products that give IT users compelling new features–and P2P doesn’t measure up. Want further proof? The P2P Computing & Development conference scheduled to be held in New York next week was canceled due to lack of interest.

I’ll be keeping an eye out for useful P2P products, and I’ll let you know if I see anything that makes me change my mind–but right now, P2P in business is mostly hype.

Ditch your desktops!

Despite two extended stints with a snowblower, my driveway is still covered by a thin layer of powdery white snow. I spent the first two days of last week working at home because the governor of our fair commonwealth announced a state of emergency ahead of what forecasters predicted could be a nor’easter to rival the famous Blizzard of ’78. (It didn’t, by the way.)

If you’re reading this in your office in Key West right now, you might have very little sympathy for me. But we all have problems that interrupt the normal flow of business, be they as grand as Hurricane Hattie or as mundane as a child’s illness.

Luckily, we can still be productive when we’re at home. Where our parents might have had a spare bedroom, we now maintain a home office. Broadband connectivity options give us links almost as fast as those we have in the office

But the one thing that stops many of us from making our homes an office away from the office is our computers. Many of us are tethered to our desktops by the little off-white mini-tower at our feet. We’ve made our personal computers truly personal, with all our e-mail, data files, and customized desktop settings and macro programs. Even if we have a PC at home, it lacks some of the essential tools of our trade.

That’s why your next work computer, and that of virtually all your company’s staff, should be a notebook. You and they should be able to unplug your PC and take it with you.

The enhanced productivity of newly mobile employees is obviously the pro in this scenario. What are the cons?

Well, notebooks cost more than desktop PCs of equivalent power. For example, a Sony VAIO PCV-J120 desktop with a 700MHz Pentium III CPU, 128MB of RAM, and 20GB hard drive costs about $850, whereas a comparably configured Sony VAIO PCG-SR17K notebook costs about $2,200. If you need to economize, you won’t get noticeably lower performance from a less expensive computer with a slightly slower CPU and a smaller hard drive.

Either way, it doesn’t take too many productive days by employees who would otherwise have lost the time to make up that difference. Let’s say your company takes in $200,000 per employee per year–a reasonable figure. You lose $800 of productivity per employee per unanticipated day out of the office.

Yes, notebooks are more fragile and more easily lost. There’s not much you can do about that, but be sure your hardware is insured before you let it out of the building, and put encryption in place on any machines that host sensitive data.

I don’t see a lot more downside. Thanks to USB and PC Card peripherals, notebooks are as expandable as desktops. If you’ve run into any problems going this route, please share them.

Of course, there are users for whom a notebook just won’t work. Artists and engineers who need the biggest screens available aren’t going to get that in a portable computer–though an external monitor can solve that problem. And those who need the very latest CPUs won’t find them in mobile platforms, or if they do, may find the costs prohibitive.

But for most business users, desktop systems are now dinosaurs. In the name of competitive advantage and minimized downtime, now is the time to outfit your workforce with nothing but notebooks.

Don’t buy software without support

When you buy software, you should get support and documentation for the price you pay. Period.

Once upon a time no vendor would think of releasing a product without at least an initial period of free support and two or three manuals. Often support was available at a toll-free number. Many companies, to their credit, still offer free support. Macromedia, for example, gives registered users 90 days of free support. But other vendors have turned support and documentation into profit centers at the expense of their customers.

Take Microsoft. When business users buy Microsoft Server or Advanced Server, how much support do they get? Zero. If they can’t solve a problem by wading through Microsoft’s free online KnowledgeBase, they have to buy a Microsoft Authorized Premier Support contract for $9,900 for 15 incidents, or $660 per incident. For the same price, Microsoft could fly an engineer out to fix the problem!

You need support because problems are most likely to crop up when you first install and begin to work with a product. Documentation, if provided, can be a big help in this process. A reference guide can be an invaluable resource when a product exhibits unexpected behavior. I like printed documentation, but if the vendor only sends a CD-ROM or maintains documentation on its Web site, at least I know where to turn.

How much documentation does Microsoft bundle with its business products? Very little. Instead, Microsoft has set up a whole division, Microsoft Press, to print the manuals that should be included with its products and sell them like books.

If you’re an investor, you have to admire the way Microsoft has leveraged its market-leading position into forcing users to pay more for its software than the price on the invoice. If you’re a customer, however, you have to wonder if there’s a better way. By the way, Microsoft isn’t alone in this practice. Other enterprise vendors, including Novell and Check Point, also offer only paid support.

I don’t have a problem with companies charging for value-added support services. Maybe you get the first 30 days free, then buy a yearly subscription — but the subscription should be a reasonable fraction of the cost of the software itself. Maybe you get the first three incidents for free, then pay a fee for additional ones. That won’t bankrupt any vendor, and it makes a huge difference to the customer’s experience. If the software is set up right, chances are you’re going to be fairly satisfied. If you can’t get it going without paying for service above and beyond what you paid for the software, you have a right to be upset.

Microsoft’s support contract is a major hidden contributor to its products’ high total cost of ownership at a time when the competitive environment is better for customers than it has been for more than a decade. Linux is a viable operating system for Intel clients and servers, and many third-party applications can read Microsoft Office-format files. If enough customers vote with their pocketbooks, Microsoft — and other miserly software vendors — may be forced to renew their focus on giving customers what they demand.