Robert Pogson

Last fall, Rob Enderle, who consults for M$, wrote “Are Thin Clients a Bad Joke?” on www.ITBusinessedge.com. He makes a few points which matter to the very largest organizations, perhaps, but not to most of us who work for ourselves or smaller outfits.

He wrote:

• “ Vance seems to have stepped back into the late 90s and spent time reading too much old Larry Ellison and Scott McNealy collateral from that time to get so excited about the Samsung PC-less monitor he is hyping.” – starting with an ad-hominem attack is not a sign of a strong position
• “Servers are built to handle large amounts of data but not designed to multiplex massive numbers of users pounding on the processor and internal systems all at once. They are data boxes, not processing powerhouses. They tend to bottleneck really badly, virtualization or not, because they simply can’t cost-effectively do the processing for lots and lots of people.” – he must be writing about M$’s idea of a server. I have seen dual core servers run thousands of proceses with snap in GNU/Linux and it is much more efficient to have servers working hard than quad-core systems idling all over the place
• “Sixteen-way quad cores (and in this kind of implementation, you can really never have enough cores) remain relatively low volume and are comparatively very costly. In addition, data centers are coming close to hitting or exceeding thermal limits. If you started dropping several hundred cores into most of them, you could likely have a BBQ in the room just by bringing your meat (or your body) into it.” – this may be his best point but it mostly applies to the big guys who have been around a long time and filled their server rooms. Most of us have one or two servers that run the whole show. Even then, the total power consumption of a thin client/server system is less than thick client/server. He does go on to mention that most folks use file servers anyway so there is already a file server idling. It might as well work hard. Terminal servers for smaller operations can run right on the file server giving RAID benefits and no network lag.
• “Finally, what made the PC industry go were standards and, for thin clients, there are far too few of them. This means if a company does make a massive investment, and it is generally massive, in a thin client platform they are stuck with the vendor.” – again, he is writing about that other OS. GNU/Linux and other unices have good old X, a fine standard. It can be forwarded over SSH for security and optimized by NX, but it still works. Most users find GNU/Linux thick clients pretty snappy. GNU/Linux thin clients work the same way except there is a network link between server and client on the display/mouse/keyboard. Thinking small does something to “massive”. Where there are a hundred thick clients one can insert one or two terminal servers and 100 thin clients for far less cost than 100 thick clients and the thin clients do not need upgrading every few years. e.g. PXE-booting thin clients are about $100/box and a terminal server to run 100 of them is about $3000 or less for hardware. Schools have packaged distros designed to set one up so there is scarcely more manpower required than installing one thick client and far less power and material/freight
• writing about mobility, ”
  The problem with thin client computing is that if you aren’t connected, you basically have a cute little tech brick on your desk.” – again, why raise the irrelevant issue? No one would make the mistake of replacing a multimedia workstation or a notebook used off-premises with thin clients. Thin clients work for the majority who still sit at a desk. If you move from one desk to another, with NX, you can connect to your session like you never left it. Few of us compute while actually walking…
• “for any new technology to succeed, it generally must embrace the technology it replaces first.” – a play right out of M$. Disruptive technology can do the end-around. There is no need to go up the middle. Using thin clients with that other OS gives you some of the benefits of using thin clients but not all. You need twice as many servers because that other OS lacks shared memory. If 100 users are running a browser, it is silly to have 100 copies of the executable in RAM. Rob must be thinking we like paying per-seat/per-server licences that bring us no benefit except the privilege of repeating the process. Not using shared memory ramps up the number of servers we need.

Of course, the premise of Rob’s fine article is wrong. Thin clients are doing well. Although they only make up about 2% of unit-sales of PCs today, they last about 3 times longer than a PC because there are few moving parts (the fanless ones) and they are not vacuum cleaners. That means about 6% of PCs are thin clients. The benefits are real and appealing. Growth is several times the growth of PCs even before the downturn in the global economy and this growth will continue. So, Rob is spreading FUD for M$. While thin clients are not an M$-killer they are one of the thousand cuts (many self-inflicted) that M$’s OS is experiencing. Thin clients work and work well for most of us. My experience has been that they cost half as much as thick clients to install and several times less to operate. Then there are the obvious advantages that appeal to end-users:

• small footprint
• low noise
• low heat
• greater up-time
• more seats can be added cheaply so there may be enough

Thin clients are a disruptive technology like netbooks. Unlike netbooks, they are not usually a consumer item but that could happen with the multi-pc home. These disruptive technologies are like mosquitoes. They just keep coming. They are a threat to M$ because that other OS is poorly designed and M$’s licensing system and welcome-mat for malware is too horrible to continue.