Where Are the Cheap Thin Clients?

Darren Ginter writes "I find many aspects of desktop virtualization compelling, with one exception: the cost of the thin clients, which typically exceeds that of a traditional box. I understand all of the benefits of desktop virtualization (and the downsides, thanks) but I'm very hung up on spending more for less. While there are some sub-$200 products out there, they all seem to cut corners (give me non-vaporware that will drive a 22" LCD at full resolution). I can PXE boot a homebrew Atom-based thin client for $130, but I'd prefer to be able to buy something assembled. Am I missing something here?"

Comments

Yes

[Anonymous Coward]

Am I missing something here?

A cheap thin client?

*ducks*

Re:Yes

[Anonymous Coward]

In my experience ALL clients are cheap.

Re:Yes

[orlanz]

And few consultants are thin.

Re:Yes

[NotBornYesterday]

Which color? Amber, green or white?

It's like bicycles...

[Eric Smith]

The more you pay, the less you get.

Though not for the same reason. You get a complete PC for less than a thin client because complete PCs are made in insanely high volumes compared to thin clients, which are a niche item.

Re:It's like bicycles...

[geekmux]

The more you pay, the less you get.

Though not for the same reason. You get a complete PC for less than a thin client because complete PCs are made in insanely high volumes compared to thin clients, which are a niche item.

Er, sorry. I consider dual-socket desktops with 64GB of RAM and 8 cores attached to a 30" monitor running 3D CAD programs a "niche" item. Thin-client hardware has been around now for at least 10 years. I'm struggling to find the connection there, especially when those that truly find the value in deploying this hardware usually do so with an order for hardware in the hundreds or thousands.

They charge what they want to charge more likely because companies like WYSE know that when you buy their hardware, the functional lifespan is likely 2 to 3 times that of a traditional desktop, and it's gonna be a while before you're knocking on their door for a purchase again.

Re:It's like bicycles...

[Znork]

It is with the bottom end desktop boxes

Which usually contain COTS hardware. Which a thin client can also use today. Like the submitter said, slapping together a thin client is easy.

they have to sell them at a higher price to recoup development costs

There are basically no development costs in this case, nor are the components high margin enough that production volume can make a significant difference in purchasing price. We're not talking special hardware here, we're talking miniITX/laptop MB's which are produced in the bazillions range whether or not a thin client producer uses them.

Personally I'd say the higher price is because the target market is almost fully corporate and corporate purchasers usually have difficulty comparing prices with anything that's not explicitly listed as equivalent. Which gets you the old triple-the-list-price and then let them negotiate a 50% discount and the customer will feel good about his leet bargaining skillz.

Re:It's like bicycles...

[drinkypoo]

The more you pay for a bicycle the less you get? What on Earth are you talking about?

The most expensive bicycles are not full-suspension mountain bikes that can go anywhere. They are racing bikes with few gears and really suited only for flat terrain (although with a sufficiently powerful "motor" you can go up hills, of course.) So yes, the more you pay, the less you get. It's not just weight, but it's largely because of weight.

Re:It's like bicycles...

[vegiVamp]

> The more you pay for a bicycle the less you get? What on Earth are you talking about?

It works exactly the same for bikinis, really.

Re:It's like bicycles...

[Daengbo]

People are looking in the wrong place. Move outside Intel and AMD.

$99 for a computer in a keyboard from Norhtec. (In fact, the prototype is still linked in my sig, but I have no connection to the company.) Video is available at Linux For Devices, but the Gecko Surfboard doesn't appear to be listed on tNorhtec's site yet.

Re:It's like bicycles...

[dimeglio]

I think the savings in deployment and long term maintenance of these terminal units are just an illusion. 1. it simply switches the cost of the workstation maintenance to the back office as you need an immensely powerful data centre to drive thousands/tens of thousands of these terminals; 2. you still need a service desk as most requests we get are for new employee accounts and handling typical release incident; 3. people want to stay competitive and having a one size fits all typically prohibits one-offs, even if there is an obvious advantage; 4. problems affecting a cluster will affect everyone so you still need backup PCs for critical service delivery. It was a great idea when all you needed was amber on black text and typically only ran 3-4 applications but now people expect full multi-media experiences, web 2.0, and although terminals are able to a certain degree to deliver these, it is often awkward and demands more than a cheap disk-less unit.

Re:It's like bicycles...

[kFiddle]

This basically just reaffirms the submitter's point. The PC-in-a-keyboard is not a thin client--it's a full, although lightweight, computer in a keyboard. It's $100. Want to buy an actual thin client? Expect to pay $300-$1000. Throw in a keyboard and monitor, and that ups the price quite a bit.

Also, the argument that thin clients are "specialty" items that drives up production costs doesn't hold up, since one would assume the $99 computer-in-a-keyboard is also a specialty item. It contains, at a minimum, a hard drive and a keyboard, which is already much more than a thin client has (not incl thin client laptops).

So why are thin clients so expensive? I've had the same question for a while now, since I've been looking around for a thin client laptop that's cheaper than a traditional laptop/netbook. So far I haven't succeeded, with most thin client laptops being much more expensive.

My guess is that the marketers hear phrases like "high security," "low energy consumption," "remotely managed," "longer longevity," "virtualization," "cloud computing," etc and think they have features that can drive the price up. The geeks, though, understand that they could build their own "thin client" by just subtracting physical parts from their existing computer and doing a little configuration.

Re:It's like bicycles...

[Sikmaz]

Let me answer your points directly as someone who has been doing some POC's of thin clients in a large (40k+ environment)

1. it simply switches the cost of the workstation maintenance to the back office as you need an immensely powerful data centre to drive thousands/tens of thousands of these terminals;
True except it is always cheaper to manage and maintain those systems than desktops. We know per unit how much each desktop costs us to manage and maintain and we also know the same information for our big-iron boxes and Citrix farm and it came out that if we could serve 20 users per server it was a large cost savings and it helped with support. We even got savings at 10 per

2. you still need a service desk as most requests we get are for new employee accounts and handling typical release incident;
You need this now anyway in a large enterprise environment and you now need less deskside people and remote support is easier.

3. people want to stay competitive and having a one size fits all typically prohibits one-offs, even if there is an obvious advantage;
Not if you do VDI which means you deliver a full desktop to the users

4. problems affecting a cluster will affect everyone so you still need backup PCs for critical service delivery.
No you just have a multiple deployments and redundancy. In most large corporations most apps are client server (Regardless of if that is a fat client or web client) so there is experience in making systems redundant.

Does it work for every user? No but it does for most, the challenges are:
1) The initial cost of deployment
2) User and business acceptance

If you can solve those issues you will experience year to year cost reductions.

Nettops?

[sznupi]

http://en.wikipedia.org/wiki/Nettop

Comes assembled, quite cheap, can drive usual resolutions, often Atom/x86 compatibility...typically has few redundant things though, like HDD; but that might be useful, together with x86, in case you change your mind.

Re:other costs

[Anonymous Coward]

A PXE-booted Atom board has neither drives nor fans.

Cart before the horse? Use a PC if its cheaper!

[Anonymous Coward]

To me the compelling aspect of virtualisation for the desktop is to be able to use a standard computer to access specialised systems, such as CAD (check out RHEV with SPICE), legacy software or test environments. At work our conference room PC's are actually normal PC's that connect to a 'conference' room virtual machine, it allows instant display of said specialised systems without effort.

AB

1996 called,

[type40]

they want their "future of desktop computing" back.
Seriously, I remember talking with some IBM engineers back in high-school and they were so certain that thin clients were the hot new thing that would change the face of computing.

You want to know where to buy thin clients? Goto www.dell.com and buy the cheapest POS they have with a fast network card. Thin clients will always be a more expensive niche player to the PC. After all what is a thin client? A PC with no local storage that can only work if it has a network connection.

Re:1996 called,

[timmarhy]

err no they aren't. you can buy an entry level thin client for $99.

thin clients never caught on because not enough MCSE's get taught about them and the CIO doesn't like all the restrictions it puts on his playing of porn.

in environment's where things like CAD are used thin clients aren't a viable option (yet), but for a lot of businesses it's by far the best way to go.

Re:1996 called,

[Keruo]

The problem with thin clients isn't the lack of knowledge. It's the break-even point.
In order for thin clients to become more affordable than deploying standalone workstations, you need to deploy atleast 200 of them, and 200 workstations rules out a lot of businesses.
Cost of licensing and server infrastructure is really the problem, not the cost of thin clients themselves.

Re:1996 called,

[gedhrel]

The organisation I work at (it's a university) spends about a million quid a year because people fail to turn off PCs overnight. The running costs of your cheap Dell POS are much higher; the power consumption too.

For clerical and administrative staff, we can put 7-14 virtualised desktops onto a single box/blade - more with non-whole-stack virtualisation or terminal services. We put our heat generation in a few places, we do get better utilisation. We also export pictures of our data to users, not the data itself, which is quite a bonus.

The downsides are what you'd expect: mostly, we have fewer spindles to deliver storage to the desktops (this is the biggest issue we face, I think); multimedia is okay-ish; for heavy computational users there aren't really gains to be had.

It's certainly got its place. Anyone selling you a "one size fits all" for your organisation probably doesn't understand your organisation, but this isn't not a completely incredible approach.

Re:1996 called,

[Clover_Kicker]

Or in other words, the IT becomes "single point of failure." While PC can run without any infrastructure or IT involvement.

Eh?

Without email, files on the network, network printers, whatever corporate apps you have, internet... what good is that PC?

Market Segmentation

[RAMMS EIN]

I think the explanation may be market segmentation. Thin clients are aimed at large organizations, where a few hundred dollars for a machine is chump change. They will happily buy greatly overpriced thin clients, because even the cost of an overpriced thin client on a desk is still dwarfed by the cost of the employee at the desk.

For home users, the picture is different, because they tend to see the computer in isolation. But the vast majority of home users wouldn't want to buy a thin client at any price, because they wouldn't know what to do with it.

If you want a cheap thin client, I would recommend to either buy one second hand (you can get them for under 100 dollars), or to just get whatever box you can and pretend it's a thin client.

Sun Ray's work well and are cheap

[therus121]

Have a look at the 'Sun Rays' from Sun - they've been around for years; they are cheap and very reliable: http://www.sun.com/software/index.jsp?cat=Desktop&subcat=Sun%20Ray%20Clients The prices shown on the Sun site are list-price - we get a Very healthy discount off of this, which brings the prices down even further.

Re:Sun Ray's work well and are cheap

[up4fun]

I agree with this, too. The sunrays are excellent.

But I think the OP is still missing something here. The transition to VDI is not just about replacing one box with another doing the same old same old. It is also an opportunity to start to transition away from local storage, login, screen savers, etc. While there are many many advantages at the back end, there are also some significant gains at the front-end, too.

As an example, the sun rays have card readers that allow you to authenticate to the back-end very quickly. Using this feature you can roll out always-on desktops that let your users sit down at a desk, any desk, pop their card in and get their desktop, just as they left it, anywhere. As they get up, their card goes with them. No need for screen savers and the whole thing is very very fast. This kind of facility is a big win for our users. No more logins! No more password resets!

So perhaps consider VDI as a way to seriously improve the end-user experience of computing.

D

Re:You're not missing anything

[OolimPhon]

Eh? Don't know what you're on about. I'm sitting here posting this on a thin client. I have a standard PC keyboard, monitor and mouse. I am looking at a normal PC session (in my case Gnome on Linux, but whatever). No retraining required, either for software or hardware. My hardware is an old PC with nothing but the motherboard left in it, running LTSP client. Cost me effectively nothing.

Thinnish thick clients.

[SharpFang]

A minor computer firm is subcontracted to assemble cheapest PCs. They build normal self-contained PCs running the cheapest OEM Windows available. These are $80(+OS) machines running on parts that are a storage surplus after they went out of sale. Then they install the "thin client" software which is some kind of Telnet or VNC or a web browser with intranet connection, pointed at a PHP web app.

So basically the employee boots up the computer normally, starts the app fullscreen and does most of the work remotely.

This has several advantages. The workstations can be troubleshooted locally. They can back up your work if network connection goes down. They allow for custom PC hardware (card readers, barcode scanners, webcams for teleconferencing and so on). They can be upgraded if the need arises, and fixed using off-the-shelf hardware (unless it went so obsolete it's unobtainable). And due to economy of scale, they are cheaper than dedicated thin clients despite being way overpowered.

I've seen quite a few markets and institutions running a system like this.

CapEx vs OpEx

[Krokant]

Don't forget that the biggest cost in a client is not necessarily the purchasing of the hardware (which is obviously the most visibile cost). Various studies (Gartner, IDC, ...) indicate that a PC that is purchased for $500 (one-time cost) in fact costs somewhere between $1500 and $4500 per year (!) to manage. These hidden costs are mainly into the backend infrastructure supporting these PC's in corporate environments, people managing them, deploying software on them, ... Google for desktop TCO and you'll find plenty of information. Sure, you might disagree with the exact numbers provided by a Gartner /IDC /Forrester but at least it gives an indication.

For thin clients (and desktop virtualization for that matter), this is also where the cost savings are. No serious VDI vendor will tell you that the CapEx (investment in hardware, licenses,...) is cheaper with thin clients and virtual desktops: you need to buy additional licenses, you're going to run desktops on server hardware (ok, 100 at a time on the same box) and then I still didn't start about the licensing galore (Microsoft VECD, Citrix XenDesktop or VMware View or...). The real cost savings are in the fact that it's much easier to manage, and being able to let your very expensive system administators do something else than troubleshooting a desktop (which costs you twice for the end-user downtime and the sysadmin troubleshooting it).

The same goes for thin clients: the up-front investment is larger, but they are very easy to manage (plug into the network and the thing autoconfigures itself, pointing you to your virtual desktop -- which means fewer expensive sysadmin interventions on-site for replacing hardware!), they live longer compared to traditional desktops (these used to have three-year lifecycles whereas thin clients typically have a five-year lifecycle -- roughly speaking you'll need to buy two traditional desktops for one thin client in a 5-year desktop lifespan; I'll concur to the fact that with the economic situation, you'll see prolongued lifetimes for both thin clients & desktops but the idea remains the same, numbers might differ today).

So is the thin client cheaper? In most situations and looking at the total picture, sure it is. Even despite a higher up-front investment. The real problem is not really the price of a thin client but whether your applications and IT environment support thin clients/server based computing (TS/Citrix/VDI).

Sidenote: I work for a consulting firm where I work a lot with VDI & Server Based Computing in general; we strive to be independent as possible (trying to nuance the vendor claims as much as possible for our clients) but that might mean I am a bit biased towards using SBC if it works ;)

That's because thin clients are PCs...

[itsdapead]

"I find many aspects of desktop virtualization compelling, with one exception: the cost of the thin clients, which typically exceeds that of a traditional box.

Thing is, if you're using office productivity apps or database front ends (the usual applications for desktop virtualization) then the most computationally intensive part of the job is probably rendering the user interface - so your thin client needs to have pretty much the same CPU and GPU clout as the desktop it is replacing. The Flash RAM costs as much as 10x the amount of HD storage and (since most people expect Thin Clients to be Thin) you're probably paying a premium for laptop-class components. The only real saving is DRAM - which is dirt cheap.

Also, since the main market for these is corporate, any retail prices you see will be inflated so that corporate clients can be offered a nice "discount".

Acer Aspire Revo AR1600-U910H

[_LORAX_]

http://www.newegg.com/Product/Product.aspx?Item=N82E16883103228

$200 and should drive a 22" monitor no problems, can't confirm PXE bootable, but with 160GB HD it should be easy enough to load up a netboot stack.

Re:I'm Confused

[larien]

There are varied benefits, but some highlights:

• Desktop breaks? Ship out a new box, they plug it in and away they go. You don't need to worry about what software they need as it's all on the server.
• Security - no hard drives on desktop which can be stolen.
• Patching/maintenance. Would you rather maintain patches on 1000 desktops or 10 big boxes in the data centre?
• Power/cooling/noise at sites. A "real" thin client (as opposed to a PC masquerading as a thin client) will have minimal power requirements which leads to less cooling and noise (no fans or crunching hard drives)
• Portability. I don't care which desk I sit at, my virtual desktop will automatically have all my apps. If you have a solution like Sun's Sunray, you can even log out of your Sunray half way through writing a document, move to another desk (possibly in another city) and pick up the doc where you left off.
• High bandwidth apps run in the same data centre as the database server/whatever and you only get the screen updates down the wire which can be more efficient.

We build ours

[markdavis]

We use over 150 "thin clients" on our network, all Linux based and all controlled by a single (large) Linux [xdm] server. We used to use "real" thin clients (Xterminals) by Tektronix, but as their prices rose and the price of cheap, fanless, low power, small, VIA boards dropped 8-9 years ago, we decided to start making our own.

We have not regretted the decision. Now we have complete control over the hardware and software. We have the ability to run real local clients when necessary.

Right now, we are in the process of upgrading to fanless Atom 270 based motherboards from Jetway. Total cost- about $250/ea.

About your hangup...

[Eric S. Smith]

I'm very hung up on spending more for less.

Stay away from "enterprise solutions," then — or, rather, make very careful comparisons between the cost of buying a ready-made thing and a DIY effort.

Am I missing something here?

That the thin clients you've been looking at are priced for fat organizations (with, possibly, thick decision-makers).

HP Thinclients

[nukem996]

HP has their own Debian Linux based client client OS called ThinPro. If you want to add more packages all you have to do is add the standard Debian repo's to /etc/apt/sources.list and your good to go. They're pretty flexible if you know some basic Linux. The best part is they have a much fuller Linux base then many other Linux thin clients. They support even more advanced features such as multimedia redirection(video and USB) as well as the basic XDM, ICA, RDP connections. All of them can drive almost any monitor from a standard 17" LCD to dual 30" LCDs. The cheapest model is ARM based. Its basically a Marvell OpenRD or Netplug with a video card and smaller disk space. All the others are x86 based and vary in speed and price.

Options

[rwa2]

They're getting there, just be patient!

I'm about the evaluate the Fit-PC2 for work, which can be had in diskless forms for under $250. http://www.fit-pc.com/

And I'm currently posting from an EeePC 901 running eeebuntu, which is actually quite a bit better and can be had for under $200. Plug in an external monitor, and rig up the built-in LCD and peripherals as a fancy KVM switching interface for your various VNC, RDP, VMware, NX, etc. backends. I'm really impressed by the Compiz desktop performance, so you can still get pretty slick transitions between various sessions on different virtual desktops.

And I'm really looking forward to the explosion of new nVidia ION netbooks and nettops, which will actually give a real nVidia 9400 GPU and dual-core Atom processors to these "thin clients", which means they can actually be used more or less like a real box in terms of running web-based interfaces and things without stuttering and pausing occasionally.

So with a dirt-cheap nettop, unfortunately you'll pay a little bit more than your target, but at least you get extra features (like a small SSD, built-in speakers, keyboard/mouse/multitouchpad, and maybe even a webcam, etc. that you could probably put to good use with a bit of creativity.

There are significant savings to be made.

[Colin Smith]

I think the savings in deployment and long term maintenance of these terminal units are just an illusion.

1: Unix/Linux systems[10] use copy on write. You load an application or library once and use it for the many users who are running the same application. The application runs significantly faster because the CPU cache and even more significantly, disk I/O cache hit rates are far higher than on a desktop system which is running half a dozen unrelated apps. This means you don't need 1000 servers to handle the load of 1000 desktops, or even 100. Your system utilisation goes from ~3% to ~90%.

Desktops. No maintenance. No 3 year upgrade cycle. The money can be spent adding business value instead.

Your desktop support problems switch from a linearly increasing management headache to the logarithmically increasing infrastructure management headache which you already have anyway.

2: You need a service desk anyway. You don't however need a desktop support guy for every floor, or local mail and file servers with the additional storage and management cost that implies. With a centralised infrastructure, distributed filesystems like AFS actually make sense, and can reduce or eliminate data duplication and duplication of business processes.

3: In what way is a remote desktop one size fits all? 95% of business users barely need more than email. Those who do need more can be provided workstations/whatever if the advantage is obvious enough.

4: You run a redundant distributed compute cluster. See Condor, GridEngine etc. The nodes are independent. Killing one, or even some of them just means others get used. You lose the network or network services? Exactly how useful is a standalone PC anyway?

although terminals are able to a certain degree to deliver these, it is often awkward and demands more than a cheap disk-less unit.

The cheap diskless units are bog standard PCs without disks. If you can stream it to a PC, you can stream it to a PC running as an X-term. ESD just isn't that difficult to set up

[10] Windows terminal servers are another matter.