Server Based Computing: Goldmine Or Snakepit? (Part 1)

If you would like read the next part in this article series please go to Server Based Computing: Goldmine Or Snakepit? (Part 2).

Introduction

This article isn’t of a very technical nature but more of a conceptual one. Wait! Don’t Alt-F4 me just yet. Most of you are probably in the Server Based Computing business one way or the other, right? Well, regardless whether you are or not, I think it pays to have a thorough knowledge about Server Based Computing and what all the fuss is about. So in this article, I will tell you more about how Server Based Computing came about, why you should or should not use it based on my own experiences in the field.

Scope

Since Server Based Computing is such a vast subject, I would like to let you know, up front, what portion of Server Based Computing I would like to discuss. I’m going to focus on the Windows “part” of Server Based Computing (Terminal Server) because this is by far the most implemented Server Based Computing option and because I have limited knowledge about the other types of Server Based Computing that are out there. Yes, there are other solutions. For example, Linux has its own Terminal Server called LTSP (Linux Terminal Server Project) and Sun does its thing with Sun Ray. Also, I’m not going to go into the whole buzz around VDI (Virtual Desktop Infrastructures) which is very interesting and also very Server Based Computing. If you, however, cannot contain yourself and want to know about VDI, you should definitely read this article by virtualization guru Ron Oglesby.

Speaking of buzz, this article will not take into account the whole myriad of virtualization techniques that has popped up within the last year or so. Although it seems like Server Based Computing and virtualization are slowly converging, I’m going to stick to the hardcore basics of Windows Server Based Computing.

History of SBC

OK, so I lied. I am going to talk about other Server Based Computing initiatives a little bit but that’s just to be historically correct. Here’s the history of Server Based Computing in a nutshell.

Way back in the day – the 1950’s – the first computers started to make their way into the larger companies to perform bulk data processing. These computers were dubbed Mainframes. Later on these mainframes could be “fed” data by users connected to the Mainframe by dumb terminals, green screens or terminal emulation. Wait a minute … what’s that word? Terminal? That’s right, you’ve just witnessed the birth of Server Based Computing. The Mainframes in conjunction with the terminals formed the first implementation of Server Based Computing.

The 1980’s hosted the boom of Personal Computing (as opposed to Server Based Computing) and spawned PC mania (which continues today with me writing this article on one). With the dominance of the PC, Client-Server computing flourished. However, in the early 1990’s there was a company which licensed the Windows NT 3.51 code to make the OS multi-user capable. This company was (and is still) called Citrix.

Citrix released Winframe in 1995 and it was quite a hit. The success of this multi-user version of Windows made Microsoft license the code “back” form Citrix in 1997. Microsoft used the code to make their Server version of Windows multi-user capable in the form of a Windows component. After an initial slow start (Windows NT 4.0 Terminal Server edition), Terminal Server and Citrix started making up a significant portion of computing infrastructures in all kinds of networks. The bursting of the dot-com bubble also initiated a demand for cost-effective computing, a bill which Server Based Computing fits nicely.

Today with Windows Server 2003 ( Longhorn on the way) and Citrix Presentation Server 4.0, Windows Server Based Computing is a valued and mature concept in computing.

Why Do Server Based Computing?

I touched on the main reason people step into Server Based Computing in the previous paragraph about cost-effective computing. It’s all about the money. People want to run their computing infrastructure as effectively as possible and yet spend as little money as possible. This makes total sense since most computing processes are there to support the core business process, not be the business process itself.

Server Based Computing allows for very cost-effective computing and, as you will see throughout the rest of this article, many of the advantages of Server Based Computing are related to cost. If you watch the market a little bit then you’ll soon get bombarded with ROI, TCO and more similar terms. I’ll try to steer clear of these terms and just tell you what’s what.

Pros

So, what are the actual advantages of Server Based Computing? Ok, here goes:

  • Savings in Desktop Management
    Because you have Thin Clients instead of PCs / Fat Client you can save a lot of money for Desktop Management. No more deploying hotfixes to desktops, no more (re-)installing applications or even operating systems. No more pesky users messing up their PCs. And there are more examples. The point is that if you can get rid of those PCs (or turn them into Thin Clients) then you can save 50% on helpdesk/management costs.

  • Hardware savings with the use of Thin Clients
    When you utilize Server Based Computing, all you need is a computer with a client that can connect to the Terminal Server. These computers do not need a lot of resources (processor, memory, disk). Hence the name Thin Client, and hence the acronym for Server Based Computing: Thin (Client) Computing. Thin Clients (should) cost a lot less than a conventional PC (in the Server Based Computing world a PC is called a Fat Client).

  • Savings in Power Costs
    Huh, power costs? Yes, power costs. If you utilize Server Based Computing properly and the majority of your clients are thin then you can save a lot on your electricity bill. I do not have exact figures but the average Desktop has a 380W power supply while some Thin Clients have a 5W power supply. You do the math.

  • No Local Data
    True Thin Clients do not have local storage. I mean they do not have a local disk. This seemingly small detail rids you of a lot of concerns. No more fear of viruses or Trojans on that device. No more fear of a vengeful employee exporting the company CRM database to his local disk and taking it to your competition. No more files scattered throughout fileservers and local desktops. Security officers love Server Based Computing.

  • Better hardware utilization
    It’s not uncommon or unhealthy for a Terminal server to have 70% of it’s memory and processing power in use all of the time. This is unlike a PC which, in comparison, barely uses all of the resources it has at its disposal.

  • Client OS independence
    This isn’t a huge advantage but it is one nonetheless. If you cannot manage to have true Thin Clients, then you can still use your old PCs to connect to your Terminal Servers. If you use Citrix then you need the ICA client which is available for almost every OS in this part of the galaxy. Even if you just use Terminal Server there’s also a Terminal Server client for Macs and even for Linux (although not supported by Microsoft). The point is: you do not need to have a Windows OS installed on every client.

  • Typically Server Based Computing has better performance
    I know this is a dangerous one. Let me elaborate as to why I say this. Servers tend to be better equipped (high-end) than PCs: they have better network connections, they operate under better environments, they have more processing power, more memory and so on. Because in Server Based Computing all the “real” computing is done at the Server (the back-end), you harness the full power of this back-end.

  • Easier Disaster Recovery
    Since the clients in Server Based Computing environments are thin and hence need no installing or configuring, the client part of any disaster recovery scenario is greatly reduced.

  • Easy Remote Access
    Since all applications are hosted in one place, remote access to those applications can be implemented relatively easy. All you need is the appropriate client and a secure transport and you can connect to the same Terminal Server as in the office and use your applications in exactly the same way. No more connection to desktops or things like that.

Conclusion

Of course this is not the end of the story. Server Based Computing isn’t a miracle cure for all your computing needs. In part 2 I will discuss some of the disadvantages of Server Based Computing and what you need to know to make your Server Based Computing initiative a success.

If you would like read the next part in this article series please go to Server Based Computing: Goldmine Or Snakepit? (Part 2).

About The Author

Leave a Comment

Your email address will not be published. Required fields are marked *

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Scroll to Top