This was a thread I started under the Software forum where it didn't get addressed. I am hoping that this will net more attention.

I work in a Council for a London borough. It is a large organisation, with several thousand staff, many of whom are deployed off campus, or engage in remote access working. Like many such organisations, the Council's software architecture is Microsoft based, and I recently proposed to the Chief Executive that we consider switching to open source as well as to thin client networking. This was a little impertinent of me, as I have no connection to the ICT department. Anyway, my argument was that this would boost our efficiencies in terms of TCO (mostly being able to resign from the habitual hardware upgrade cycles, and let go of exorbitant licensing fees) as well as would yield better productivity without having to put up with the frequent MS Word crashes, Explorer hangs and Outlook Exchange server disconnections (the Council is using NT4.0).
Not surprisingly, the ICT department want to cover their arse, so I have received an email from the "infrastructure manager" informing me that he will send me the strategy documents outlining why open source would not work for this Council, and also the decisions relating to their promotion of the thin client approach, using Citrix.

I firmly believe that GNU/Linux - either by itself, or in conjunction with FreeBSD and OpenBSD (the latter especially for the servers and security features) - is capable of doing the job. It would need to be able to provide reliable desktop and office suite application environments for several thousand people (approx. 4,000), both on-site as well as Councillors who want remote access to email and office applications, and off-campus departments scattered around the borough. It would also require the ability to be applied to a thin client architecture in the near future. I was thinking that either Suse or Fedora professional (really for the support factors and for all the eye candy and integrated applications), but I am sure that Slackware or even Debian would be well suited for the job.

My questions refer then to:

(1) Can GNU/Linux scale to the extent required, and which distro is most suited for this task?
(2) What are the likely pitfalls to be encountered?
(3) What is the definitive conclusion on these TCO debates and MS-generated FUD reports?
(4) What alternatives to Citrix are there in open source land?
(5) What would be persuasive arguments to make bearing in mind that these ICT folk want to cover their arses and have expertise and experience with MS but not with open source, and hence may be rather nervous about possible loss of status and the learning curve?
(6) Are there solid open source case histories and best practice examples that could be drawn on for examples as successful alternatives to MS?
(7) The software and architecture would have to be able to work with PDAs and public-access kiosks (such as in street-based information booths and public libraries) as well as some school systems, and with these additional considerations in mind, is open source still feasible?

There will probably be more questions in the weeks and months to come as the conversations with the ICT department folk continue. Given their reluctance to make the switch (especially as they would have a vested interest in maintaining the status quo), I am unlikely to be able to convince them, but I would like to make their resistance that much harder and even more absurd to defend and maintain, if possible

1) Point you favorite browser at these search keywords: linux enterprise. 4 million+ hits on www.google.com/linux.

2) Createst pitfall: inexperience on the part of IT personnel, as well as their clients (users). There will be a steep learning curve, which can be curtailed to some extent with directed education. Perhaps there are consultants in your area who could help.

3) Vested interests fund the research. They expect favorable reports. If the reports are not favorable, they tend to ignore the results. That's human nature. It's also common business practice.

4) I'm not familiar with Citrix or what it does.

5) The cost of implementing a switch (cost of software and support) is less than that charged by MS. However, the cost of retraining must be considered. See 2) above, in re consultants.

6) Point your browser at this search phrase, in quotes: "case histories".

7) Pda yes. Public-acces kiosk probably. Communicate with school systems, even those using windows, yes. Though, Linux would probably have to be set up as the server, and windows lans in school systems as clients. Basically, Linux works with windows better than windows works with Linux.

Excellent response - thank you very much. I'll come back to this if, after doing the leg work, I have further queries. It will be useful looking at the case histories, for sure.

Cheers

Look into what happened at Barking & Dagenham Council when they publicly announced that they were switching to Linux from Windows.

You can use X11 (XDMCP) itself. Linux
desktops were made to be remotely displayed.

XDMCP is like RDP (Windows Terminal Server "full screen desktop").
Normal X11 is like ICA (Citrix "seamless windows").

You merely setup your Linux server to allow multiple users to login via
X, and the clients use the XDMCP. The desktop runs 100% on the server.

There is a project called the Linux Terminal Server Project (LTSP) that
has a canned setup for this. It runs atop of Red Hat distros.

i hope that gives you a bit more info on the Citrix vs OS and its abilities.

the #1 obstacle will be the current IT staff unwillingness to learn new skills or their fear of being replaced by a staff that knows more then windows.

TCO of linux is far less then MS products and there are plenty of true real world studies out there you can google for. All I can offer is in an extremely small (2 servers and 10 end users) setting of MS vs linux.

Under win2k server and both win2k pro and winXP Pro for desktops i was spending between 20-60hrs a week depending on how bad the viruses were that week just trying to keep the 2 servers clean, up, and doing their job. yes this was AD with the servers acting as TS servers, file servers and 1 print server. no IIS running as that is just to dangerous and insecure.

About 2 years ago i replaced all of my desktops to Linux. My man hours drooped to under 10hr a week dealing with the servers as i also got rid of one of the servers and turned one of the workstations into a print server.

I am now 100% MS free and my man hours are under 1hr a week to keep track of the desktops and servers to make sure they are clean, no rootkits are installed, and everything is doing their job. i go an average of 90-130 days before i reboot any of the systems and that is normally due to a kernel update.

So over the past few years i have cut my TCO from very time consuming and keeping me away from doing my real job of running my business to make money, all the way down to a part time hobby. I now have cron running to deal with the majority of patching and updating, i no longer have to worry about viruses or spyware, my servers STAY UP and running, and things just work much better.

FYI this is all the same hardware except one station i have brought in a new iMAC and it integrated smooth enough into my linux network.

I know that is nothing compared to an organization the size of yours, but the overall affect is the same.

FYI training on my users took less then 1hr. the GUI (KDE is my choice) is very easy for any windows end user to get accustom to so far as single click, and where and what to open to do their jobs.

I use Firefox, Thunderbird, and OpenOffice for all users and we use QuickBooks, thus the need for the iMAC, for our accounting. I just use VNC via ssh to gain remote access to my iMAC from the POS systems with zero problems. I am not so large that I need more then 1 person at a time accessing QB. When I get that large, I will figure something else out.

I've been looking for info on this announcement but can't find anything. Who reported it and when?

Thanks for the detailed response Lleb_KCir. I think that you are correct about the IT staff being the main stumbling block - that, and apparently the Council have signed a 3-year deal with MS and are looking to pilot SharePoint under contract to SilverSands. For the majority of the requirements, I think the set-up you have outlined would presumably survive scaling up to meet the demands of the local authority, and although there may be difficulties with kiosks and some of the specific software requirements they have, I suppose the only other concerns would be about interoperability with MS products being used by other agencies and authorities. For that reason, it might be worthwhile keeping a couple of MS machines around, running on top of GNU/Linux servers.
As I said in my original post, I really don't think that the powers that be would go for it, even though it would save the council a mint and hence meet any number of our cashable savings and efficiency targets. And now I notice that the correspondence from the IT dept has become mysteriously silent. It is probably time to rattle a few cages again

Bum! I can't find it. It was fairly well publicised in the tech press (maybe I got the county wrong....?) that a council in the UK announced it was moving to Linux so Microsoft swooped in with their PR and Sales people and basically gave the council Windows and Office for nothing. Massive charm offensive.

If your organisation does go for Linux, I would say that you should make the switch and then make it public - present it as a fait accompli.

he he he
I bet!!!
Well, that sounds like a good strategy, but it will take the political will to make that switch happen and that just isn't there ... yet (said optimistically!!). This is not exactly the most innovative of Councils we are talking about, I regret to admit.

Present them with figures and with the "alternative" get The Facts campaigns - ie the ones not written by Microsoft. IBM and Gartner (both enormously respected and incredibly well known) have both produced some of these. And so did Novell.

Have a natter to your beancounters, see if they will give you a rough estimate of the TCO of your current setup and work with them to see if they would consider an estimated TCO of going to Linux.

Present your facts in clear, plain English - bosses really don't like technical stuff - and load up a spare pc with something like OpenSuSE - nice looking but still professional - and show them what's there. You don't even need to get too technical, this is a prelim thing after all.

Once you get the go ahead, do things like load up a Windows pc with OpenOffice and Firefox and see if the user (we need a live box for this) is willing to use it for a week or two and document any problems or issues. In the meantime, list the apps you use and the equivalent Linux ones and list the pros and cons of switching.

Really, what you want to do is set a base for a good debate on this. If it seems worthwhile, management will then likely drag in Cap Gemini or other big firm and do a proper bit of consultancy on this. You just need to get your ducks in a row first.

Bottom line: it is doable, but it won't be quick.

d00bid00b, one good thing about linux is that it plays better with windows then windows does with linux.

so having the servers running linux the workstations do not matter what platform it is. FYI kiosking in linux is easy if you are running to a central server just use the X forwarding power that linux has. been around a lot longer then anything Windows uses as they have a hacked vs of citrix(sp?)

QUOTE from Lleb_KCir]

the #1 obstacle will be the current IT staff unwillingness to learn new skills or their fear of being replaced by a staff that knows more then windows.


Well, yes and no - i think you'll find that the techies at the sharp end will be quite receptive but their major problem is that they will require training to support it. They have experience of, and training with, Windows, most of them are MCP's/MCSE's and they will require the same sort of committment from their managers to train them in Linux/OSS. The bad news is that training courses eat money for breakfast. 10 techies at £1500 each for a 1 week course kinda negates a lot of the savings you'd make by switching from MS.


TCO of linux is far less then MS products and there are plenty of true real world studies out there you can google for. All I can offer is in an extremely small (2 servers and 10 end users) setting of MS vs linux.

Well, yes and no - TCO for the two systems once they're up and running is undoubtedly lower for Linux/OSS. The costs come from migrations and training.

Under win2k server and both win2k pro and winXP Pro for desktops i was spending between 20-60hrs a week depending on how bad the viruses were that week just trying to keep the 2 servers clean, up, and doing their job. yes this was AD with the servers acting as TS servers, file servers and 1 print server. no IIS running as that is just to dangerous and insecure.

I'd suggest that your policies could have been a bit stricter! i.e. all email attachments that could be even remotely dodgy are quarantined and users have to ask to have them released. Web policies to restrict access to sites unless they are requested and a business justification given as to why they need to access that site. etc etc. I.e. lock everything down unless there is a justifiable need to have it unlocked.


FYI training on my users took less then 1hr. the GUI (KDE is my choice) is very easy for any windows end user to get accustom to so far as single click, and where and what to open to do their jobs.

Aye! and there's the rub. 3500 users need training, integrated systems need changing (how are your e-gov initiatives coming along ? Bet they're all integrated into Exchange and how are you going to replace that?) Printer drivers need changing, document management systems need changing etc etc - it's a huge job and the bottom line is that it costs money that local government just doesn't have.

I'm the email admin at a London Borough - i'm a Linux evangelist, my managers and colleagues take the mickey out of me every time i say "well, Linux could do that easier, cheaper, faster, better etc". I think that the way to get out of the MS stranglehold is a one step at a time approach. Internet Mail gateways on Win2K machines - hmm, i could replace that with a linux box running postfix/sendmail, clamav, spamassassin etc etc - that frees up a Win2K licence, thats a saving.

Mark

Sorry, i was a bit aggressive there!

What i mean to say is that rather than go for the whole "Lets move to Linux now" argument, our best way forward is to say "well, we can replace these specific windows boxes doing these specific tasks with linux and save money."

Otherwise you get into a similar situation to the whole Munich or Barking/Dagenham thing where the project spirals out of control and it becomes easier to stay with MS - do it one step at a time, evaluate and move forward.

Mark

yes without a doubt (sorry with the change on this forum, my old account died *cry*) moving something as large as 3500 users with full support in servers and other resources is not something that can happen over night. heck even with my small little 2 servers and 10 users i took time to move things around a bit.

that is part of the migration plan though. you move things in stages, start the training in stages etc...

as for training the EU to use the new GUI. that is something that can be done very cheap and on the fly so to speak. train the "area heads", for lack of a better term not knowing your organization, enough to train their subordinates. normally that would be managers, suppers, employees. then it is finished under 1 week and cost very little in time, money, and downtime.

as for the IT staff. it is not mediatory to keep the mediatory MS Tecs. yes they are trained as MCPs or MCSE or higher, but they are not trained in Linux/Unix. It is no different then a company becoming automated. that employee is fired because his job no longer Decks, or he no longer meets the requirements for the new job. Can you keep them, sure if you want to pay to train them, some are worth it, others are not. in most cases it is just simpler to higher new people to replace several of the Decks MS techs.

example:

you have 100 MS Techs under 10 IT managers with 1 CIO.

keep the 1CIO if he is onboard with the move to linux.
keep 8 of the MS IT managers and higher 2 Linux IT managers to help with the migration.
over the time frame of the migration (in the case of 3500 EUs this migration could take as long as 3 - 5 years) switch the number of Linux vs MS IT managers.
of the 100 MS Techs fire 20 of them and higher 20 Linux techs and do the same thing with your techs as you do with the managers.

by the time the migration is over you will have a staff that is able to run and manage Linux and have also given several of your current employees the opportunity to make their skills able to be retained in their job. if they can not bring their linux skills up to par for the position they are applied to, they do not get the job.

that may sound very harsh, but that is business. if the person wanting the job does not have the skill set, then that person does not get the job.