Slashdot Mirror
Thin Client, Or Fat Client? That Is the Question
theodp writes "If virtual desktops are so great, asks Jonathan Eunice, then why isn't everyone using them? However encouraged folks are by the progress virtual desktop infrastructure (VDI) has made, and however enthused they may be about extending the wins of server virtualization over into the desktop realm, you don't see analysts and developers eating the virtual desktop dog food. And even the folks you meet from Citrix, Microsoft, Quest, VMware, and Wyse — the people selling VDI — use traditional 'fat' notebooks. So, are you using virtual desktops? Why, or why not?" I wonder how long the abbreviation VDI will stick around.
Developers won't generally use them ... as with so may computer related things these days, VDI is not about usefulness, it's about control. It makes it easy to lock employees down to a standard desktop, and provision or restore them with minimal effort. It's not necessarily a bad thing, but it's not really aimed at developers.
Employers love thin clients because they give more control over the information moved in and out of the organisation. You don't have to worry about blocking Lady Gaga CD-RW disks if the user only gets a picture of the data anyway.
But then the same limitations create a constant demand for new solutions to work around problems which should be simple. How can the PHB work on the plane? What is a switch dies and takes out sixteen users?
I have seen thin clients used successfully in a doctors office, where the integration requirements are simple. I can't see it satisfying every requirement in the engineering environment where I work.
http://michaelsmith.id.au
Having used some actual thin clients, it's not bad, though I wouldn't want to game on it.
I think the core point would be that because of the way licensing works - you have to buy the client, pay for a license for the client, a license for all software used on the client(if you're going to be legal), it ends up actually being more expensive than a bottom line, but still capable full PC.
The HD adds, what, $20-40? The licenses for our thin clients was more like $100.
I don't read AC A human right
I'm not sure what you worked on and when, but with native virtualization instructions in modern processors, there is no noticable speed difference, the biggest place where people see issues is with disk contention with a badly designed storage platform behind the virtualization solution.
Curiosity was framed; ignorance killed the cat. -- Author unknown
For banks, I think that probably ought to be required. Industries like that and places that need to tighten control of the data love VDI, as it makes it a lot harder for somebody to gain access or more worryingly leave secure data on an insecure machine.
I'm going to call BS partly on this. Most of the business world is using basic productivity software, probably Microsoft Office, with some users needing access to an accounting package or CRM. Thin clients aren't so much about up front cost as they are about reducing long term support costs. Using thin clients in an enterprise or small to medium business environment gives you a lot of benefits to the long term bottom line. From a security perspective, you cut the "attack surface" of your network very sharply - from dozens if not hundreds or even thousands of desktops that each need antivirus, security updates, administration, and security monitoring, down to a handful of servers that you can lock down pretty tightly. From a support perspective, you are no longer managing all those desktops, you are now managing a handful of servers. You have all the data for your organization where you can make sure backups are happening, and where you can keep tabs on what data is being stored and where it's stored, so you no longer have to worry about that file with a million customer social security numbers or credit card numbers sitting on someone's desktop, where you won't find out about it until after it walks out the door. Also, with a good setup, you ease the pain of patch days a fair bit, since you don't have to chase breakage across all those desktops, just across the app servers. You remove the expectation of user control because a thin client is clearly not a desktop (the "but I can do it at home, why can't I do it here" syndrome). These are damn good reasons to go to thin clients on the desktop, even if the up front costs are the same or even slightly more, and they apply to most desktop users. Only "high-performance" application demands, like CAD, and software development need fat desktops. Now, on the laptop side of things, internet connections in the field aren't something you can count on, even with mobile broadband and wifi penetration, it's not always there, and it's not always good enough. so thin clients aren't going to make much headway there for a long, long time.
How does the iPad get around the problem of over saturated hotel networks or poor 3G connectivity? What does it bring to the table that isn't the same as a thin client on a laptop? That's right, it doesn't and nothing.
The promise of thin clients has never been on upfront costs. The advantages have to do with maintaining the clients once they have been deployed. Think patches, service packs, O/S upgrades, memory upgrades, HD replacements, etc. With traditional desktops many of these changes can only be done by going to each machine individually. Additionally, thin clients make backup/restore trivial whereas trying to enforce data retention standards on desktops is always a battle. While these issues may not present themselves in a small to medium sized company, trust me when I say that with thousands of installed desktops there are hundreds of people dedicated to maintaining the hardware and managing the environment.
I'm not sure what you mean by "there aren't suitable thin client options for most businesses." Most of the actual business of say a bank, or an insurance company, or a web vendor, or just about any company that isn't a full fledged software developer comes down to a few apps that rarely require huge amounts of memory, the latest video card, or even a hard drive since most of those apps just run as a client and save data on the server anyway. In fact I can think of few businesses where thin clients shouldn't represent the majority or installed systems.
You're not answering the question. Regardless of whether it's a PC, a notebook, a netbook, a tablet, a smartphone or a CPU up your uncle's ass, the fact of the matter is that poor and oversaturated network connections like you will inevitably find in a hotel room or public WiFi is going to mean very slow screen updates, slow cursor and keyboard updates, and so on. Using a fucking touchscreen doesn't magically cure these problems. It's not like an iPad has some magical faster network connection.
The world's burning. Moped Jesus spotted on I50. Details at 11.
I've done contract work for one of the major credit card companies as a data analyst, and they require us to go through Citrix clients (among other hoops) for a very simple reason: security. There is no way for us to perform an end-run around all of their security for the simple fact that they control our environments, tools, everything. It doesn't matter what kind of system we have (Windows, Mac, Linux, bottom-of-the-line laptop or ass-kicking Alienware), we're limited by the environment that they set up for us.
They could easily pay triple what it's costing them, as long as their security measures are in place, that's what matters most.
Well aren't you a special little snowflake.
Guess what? You aren't everyone. If everyone had the same needs you do, software companies - catering to the domain needs of people, companies, conglomerates, municipalities, and governments - wouldn't exist.
Just a couple things which open source won't (doesn't) do (well/sufficiently):
* CRM, particularly for specific domains
* Human accounting
* healthcare
* Legal/trial workings
* Drafting, rendering, and animation (still one of the only domains exclusively populated by 'real' desktops running Windows)
* Pretty much any and all retail establishments which want to track merchandise
As a result, we've got software which runs only on Windows. Isn't it interesting how supply and demand works?
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
Why does EVERY FOSS guy think it is about Windows and Office? Hint: It is NEVER about Windows and Office, it is about all those funky ass "mission critical" apps that do not run on Linux full stop. From the charting and accounting software, to the payroll apps, damned near everything in most business will NOT run on Linux, and a good 80% of them I'd say have NO equivalent on Linux.
Just saying to a corporate customer "Just use Linux" is about like walking up to a graphic designer and saying "You don't need that pesky Photoshop, it isn't free as in freedom! here enjoy the Gimp". Try it and see how well it goes over. It is NEVER Windows or Office, those are easy. There are a bazillion other niche apps won't run on Linux, would cost a buttload of money to have a custom version made for Linux, and that is even if you can get the data out easily. Hint: Often you can't. Nearly every SMB and SOHO in my area relies on Quickbooks. What is there even close? GNUCash? It is about as close as Gimp is to Photoshop.
So in conclusion while I'm sure it'd be nice for Linux adoption if all people did at work was use a browser or make documents, but that is rarely the case. It is all those OTHER apps that are used every single day that bite you in the ass, NOT Windows and Office.
ACs don't waste your time replying, your posts are never seen by me.
That is why we are piloting a VDI deployment this year. I work at a community college and we are rolling out thin clients using a combo of VMWare, Unidesk, and Panologic. What we were drawn to was how easy it was to control what applications the user gets, patching the desktop and applications, improving security, faster recovery times if there is a software issue (like a virus that slips past), and lastly because we have kiosk computers, being able to have non-persistant desktops (This will save us a little money as it replaces software like deep freeze).
We are in a unique situation however where we need to give users access to multiple operating systems and applications on a given hour. In the past we would need to build our physical machine's bigger then normal so they could handle windows XP + a virtual machine for windows vista or 7. For about the same cost of the PC hardware we would have purchased for the number of PC's we selected for piloting I was able to purchase the server, software licenses, and thin clients needed to do the accomplish the same task. Our hope is that we can save soft costs when it comes to managing these computers and that the thin clients will last longer then a typical PC refresh and saving us cash in the long run (as you pointed out).
Other features that weighed on our decision was how easy it was to recover a dead machine and being able to leverage de-duplication. Currently, if our CFO's dell computer died, we would need to retrieve his computer, recover his files, image a new computer (providing a suitable computer was in stock), install any custom applications he requires, and finally place his data back where it belongs (providing whatever died is not in stock to replace). Ideally all of his data would have been backed up on the network, but being a typically user he probably saved to other locations as well (And being someone like a CFO, you can't give him the usual 'you should have done it like we told you' speech). If the thin client dies, it is literally just a walk to his office and a swap of the device. If he can't wait that long he could just go to any other thin client and log in to get his desktop. Further more, UPS devices at the desktops are no longer required. If the power goes out, the desktops are still running in our datacenter with it's own generator.
Our back-end storage has de-duplication. This combined with the cloning technology unidesk leverages has us with 25 desktops deployed and only 60 gigs of storage used. While user data will increase that amount as the system is used, we could have 1 or 100 desktop and would still only store 1 copy of the OS and 1 copy of each application installed. The savings can be significant.
We are also hoping that the thin clients will ease our way into windows 7. As a college we are constantly being pushed by students, outside companies, and our own faculty to be technologically current. There has been a huge push for us to start moving to windows 7. Unfortunately, there are still many applications that are not supported on windows 7 that some staff members require. There is also course material (some certification tests) that have not yet updated their material for windows 7. Using this system we will be able to offer both XP or Win7 at any location where there is a thin client. So if you run into a document that doesn't work right in office 2007/2010 you can quickly 'reboot' into XP and office 2003 and get your job done while you wait for the author of the document to update it. This comes at a great time because a lot of our computers are at the end of their life cycle and would not meet the requirements to run windows 7. We can provide them with a thin client with XP and all their apps, then slowly rollout windows 7 allowing them a transition period where both systems are available to them.
We know that VDI won't be the solution to all of our problems. There will always been users with notebook computers, classes that require high end video cards or lots of ram (3d modeling, etc), and other situ