Mad Dog 21/21: It’s Your Gaul
November 19, 2007 Hesh Wiener
Like Gaul before Charlemagne, office desktops can be divided in three. One nation uses PCs running a mix of generic and corporate applications. Another nation uses thin, fixed-function clients, even if some of the machines have full-fledged computers inside. The third nation uses PCs, but works with software that shifts most client functionality to a shared terminal support server. For decades, friends and enemies of these nations have made more or less the same old arguments. But now client technology is really different, and so are the end users. Some choices that were right yesterday may be wrong tomorrow.
What has undermined the old arguments in favor of one or another client philosophy is pretty much the same as what happened to Gaul once Charlemagne kicked the bucket. Basically, the problem began under Charlemagne, who simply could not control all the local warlords and landlords in his empire. As a result, he couldn’t collect the taxes he needed to build the kind of army needed to maintain order.
It was a vicious spiral. When the Charlemagne’s heirs came into power, a process that left the formerly unified empire in pieces if not shreds, they were neither able to keep the gentry in line nor defend against incursions by outsiders who coveted the land of the Franks. Power shifted to Hugh Capet and his heirs, a weak family that held title because they really weren’t much of a threat to local power. Hoping to offset this erosion of the empire, the Capets started the Crusades. The Crusades didn’t help unify Europe either, and in the end they left what had been a vast Christian empire with permanent enemies among the Muslims. Sound familiar?
Yes, it’s easy to see the parallels between the dissolution of Charlemagne’s unified empire and the loss of uniformity that once characterized client devices used in business.
Most business desktops sport PCs. But the PCs are not always independent machines tied together and hooked to servers by a network. In some cases the PCs are set up to work via a shared server using Citrix Systems‘ Remote Desktop Protocol, Microsoft‘s Windows Terminal Services, or another similar scheme. Access to corporate and external resources must be made through the terminal support server. In still other cases business does not give users PCs at all; it provides them with thin clients. The clients typically use a dedicated server for any computing that is beyond the basic presentation and input chores performed on the desktop.
One might think that the advantages of thin clients and PC clients tightly controlled by a shared server would sweep the general purpose PC off the corporate desktop. Thin clients require no support (although their servers might require updates and maintenance). They are just GUI appliances, so there’s hardly anything that can go wrong with them: Either they work or they do not. Thin clients don’t have a zillion security issues the way PCs do. And when an application gets updated, the user organization only has to get the update onto the right server; the client is not affected at all because its role never changes.
When it comes to managing or restricting the users’ access to local and remote resources, it is generally a lot easier for a company to set rules on a server and thereby manage all its thin clients than to impose order on a PC-by-PC basis. While an employee at a thin client might be able to do unauthorized or forbidden things via the corporate network, that freedom is a matter of server and network security, not an issue set inside a PC that is difficult to monitor and control.
Every PC, even one set up with a lot of security restrictions, requires local support. The hardware itself is more complex and almost always includes, for example, a disk drive that a thin client does not need, so maintenance can be more costly than for thin client, too. Users who really want to bypass security restrictions on their PCs have a much better chance of getting their way than users of thin clients. In addition, PCs, particularly Windows PCs, are subject to a flood of threats in the form of worms, viruses, and other malware. They are targets of a vast array of intrusion technologies that come at them via the Internet.
But all these arguments about smart versus brainless terminals have been made since the first time a PC showed up where a green screen once lived. The only aspect of the debate that has changed a lot is the part about malware; that’s gotten more heated every year because script kiddies and hackers keep inventing bad stuff.
The arguments can all be reduced to feuds about feudalism. Corporate organizations are for the most part run along feudal lines. Some estate is delegated to a division; part of that estate is delegated to a department; part of that last bit to a workgroup. Authority and responsibility usually get handed down as part of the estates. The alternative, found in some small (and maybe a few large) open plan, entrepreneurial companies involves adhocracy. Everyone knows their role and people just try to co-operate using common sense and courtesy. (You can see why this is not so easy in big companies.) It is, in fact, possible for a company of any size to have quite a bit of republican spirit (although feudalism is still the rule). If Europe could get from Charlemagne to something approximating democracy in only a thousand or twelve hundred years, there is a glimmer of hope that Oldmegacorp could get dot-commed and not die during the transformation.
Anyhow, the debaters in the client spat (as it’s is hardly big enough to be considered a war) have for the most part danced around the issue of infected servers. Is a Citrix server really any safer from Windows virus problems than a PC? Is a Windows-based thin client presentation server? There’s not a collection of published studies on this topic. So the whole process or argument, which ought to yield some useful results, is full of holes. And the server vulnerability issue is only of gap. Another topic the fat versus thin debaters seem to avoid when they touch on security is this: Corporations could drastically reduce their exposure to malware and still keep their desktop computers if they simply moved from Windows boxes to Macs.
The reason hardly anyone wants to explore these poorly lighted corners of the client capability arena is that the big noises about PCs versus Citrix-type client setups versus true thin clients are made by people and companies with big vested interests. While most or even all the arguments one way or another about which client is best for which user might be true, who is going to swallow whole a case for or against the PC client made by Microsoft, Oracle, Google, Hewlett-Packard, IBM, Dell, Red Hat, Apple, or any other vendor that has more hidden agendas than Windows has hidden APIs.
Whether the case for a PC on every desktop is stronger than the case for a thin client on every desktop is gospel or total hooey hardly matters. It’s just about impossible to find out what users think from what is published. All a curious party can do is see what a business thought in the past by looking at desktops to determine how the money was spent. And while that usually shows how much garden variety PCs have succeeded, it says nothing about how wise the business has been. There’s not a parallel universe where that business got thin clients (and honest budget analysis).
Currently, there are some pretty good reasons for businesses to notice that the debate about client technology is a steaming pile of horsefeathers: The gap between what corporate computers feed clients and what the outside world is sending at them is becoming a chasm.
Personnel who must use the Internet to obtain information (and as a result get wide open or pretty wide open access to the Web) are beginning to see how the stuff they bring into their offices by wire is changing. The content includes more multimedia material; Web sites with sound and video and various animation technologies in their content are growing more common every day. Not all these sites that are built catch the attention of teenagers. Some of them are presentations made by agents with factories for rent. Some are industrial companies that use cartoon or movies to illustrate their processes. Some are service companies that put their executives, or at least the telegenic ones, into little Web rarees. Sure, quite a lot of this content can be viewed on the more feature rich thin clients, but as technology moves on the codecs and other components users need to surf the Web turn out to arrive first for PCs (and maybe some game consoles), then for half-a-grand mobile phones, and maybe later if at all for thin clients.
Web media is not only changing in content, it is changing in form. Current Web designs are not built for the 4:3 screens found on most office desks; those screens are Plan B. Any modern Web design is conceived with a phi screen (aspect ratio of 16:10) in mind. Phi screens and their close cousins, wide televisions and HDTV receivers that have 16:9 aspect ratios, are the designers’ targets. Because that is where the money has moved–consumers first, business, perhaps reluctantly, second.
This doesn’t mean that users should just give up thinking about what kinds of clients to buy and instead start budgeting for PCs with Vista and wide screens. It means that the whole situation needs fresh ideas. There’s no reason thin clients cannot be a lot smarter and more versatile than the ones that vendors offered (without much success) in the past. But neither is there a good reason for users to simply swallow whole the idea that a PC with Windows brings with it an unavoidable world of woe. There are, in fact, Macs. There are, in fact, PCs that can deliver just about everything a business needs via Ubuntu, Fedora, or openSUSE Linux, among other friendly distros. These alternatives to the complicated, vulnerable, high maintenance PCs might be right for some users.
The thing that has to happen, and it is inexplicably rare in the business world, is this: Somebody in the computing department has to decide to buy five or ten alternative clients and deploy them. The cost, perhaps a grand apiece if a business is good at installation, is small compared to the potential savings if end user support costs can be reduced, if applications development costs can be cut, or if a demonstration of intellectual independence helps during price negotiations with a traditional supplier of traditional Windows PCs.
All it takes is, as Charlemagne knew, gall.