Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Upgrades

Windows-to-Linux. Large Installations Handling the Changeover? 29

Posted by Cliff from the searching-for-ideas-advice-and-previous-experience dept.
Cathal asks: "Okay. I am a senior in a large Irish university, which is a mostly Microsoft house. As a member of the college computer society, I have heard that the college bean-counters want to reduce the cost of the IT dept. The IT dept are 'thinking' of turning to Linux as a method of cutting costs and improving the service to the staff and the students. I am looking for suggestions and feedback on previous experience in similar situations, (large changeovers, support)."
"What the college supports at the moment:
  • 9000 Undergrads
  • 2500 Postgrads
  • 3000 part-time students - night classes
  • About 1500 staff and lecturers
  • Print farm to support the above, with network printers available in each of the 40 or so computer labs around the place.
  • About 25 webservers with a combination of IIS, Apache(win32) and Apache(solaris)
  • 300-400 student computers in the on-campus accomodation

College resources:

  • About 2500 desktops in the college, at the moment with msoffice on win2k
  • A collection of fileservers, and mailservers, (mostly Dell poweredges)
  • Fiber backbone, 100Mbit switches, 100Mbit to most desktops, and a 20Mbit connection to net backbone (to go to 125Mbit next year)
If the college does move to Linux, there are a reasonable number of experienced Linux admins as student in the college, running a debian cluster for the compsoc, and could provide a pool of about 10 admins that know the network and the people involved in the running of the college, and could ease the changeover. If anyone has been in a similar situation, what were the pitfalls involved, and the main difficulties in rolling it out? The college *needs* (so I have been told) to have an external contract to solve problems with a defined level of service."
This discussion has been archived. No new comments can be posted.

Windows-to-Linux. Large Installations Handling the Changeover? More

Windows-to-Linux. Large Installations Handling the Changeover?

Comments Filter:
  • Well from my little experience.. don't look to replace windows everywhere. You will need to keep some windows server's running for specialist things, as well as several servers it may not be a cost saving at all because of the extra time/effort required. My current thinking is that linux can replace 70% or so of all NT servers out there at least, and there are considerable savings possible from this.
    I don't know that it's really worthwhile changing a large user base over from what they currently use (I know that at my uni the amount of windows specific apps for lecturers and other staff is pretty phenomenal). Also you probably have a lot of lab's which require windows specific software (like those wonderful first year "this is MS Office" papers).
    Of course you could talk to Redhat, and then MS and see what MS will do to beat Redhat.. (now where's that asbestos suit?).
    This is just what I have learnt from helping at two educational institutes. Hope it helps.
    • What is it that you need those 30/100 NT servers for? What are the "specialist things" that you are talking about?

      -Peter
      • That 30% or less is all those wonderful things you happen to run into like legacy accounting software which has a server that will only run on a specific version of a specific os, and there isn't a snowball's chance of getting the data converted and finding a replacement in the one shot without major disruption. There is also a lot of stuff which works on linux, but has a reduced feature set or has bugs (openldap seems to be one of these things). I'm not saying a total conversion can't be achieved, just that you have to weigh up whether it's worth the effort for some things.

  • IBM (Score:2, Insightful)

    by mlinksva ( 1755 )
    Call up IBM [ibm.com], RedHat [redhat.ie] and local shops (quick search turned up Eolach [eolach.com] and Bayridge [bayridge.ie]). They all should have experience fighting for Microsoft-held accounts and bringing Linux into that environment, and can fulfill the external contract requirement.

    Replacing MSOffice looks like the biggest hurdle to me. StarOffice is really the only complete alternative available. Maybe StarOffice 6 will actually be good enough [techtv.com].

    • BTW, another hopeful article about StarOffice [consultingtimes.com] concerning how well it handles MSOffice documents.
    • Heh.

      According to the Eolach Open Source News Worth Knowing newsletter that I received yesterday, they just went tits up.

      • If memory serves (as it often doesn't), Star Division (the German(?) company that originally produced StarOffice), went belly-up, but gave the rights to Sun to redistribute as they wished... not sure though. :P
        • True, in part to tweak Gates' nose, but there's also their goal of running it on servers. Anyway, we normal folk sometimes benefit when billionaires fight out of spite.


          hawk, economist, and advocate of the side effects of greed and spite

  • Burn a couple dozen bootable CDs with linux, samba, and a few other goodies (ie. word processing, GIMP, etc.). Give 'em to them and tell them to try it. Show them that they can get 99% of the functionality that they currently are paying thousands for, for free.

    Just my $0.02

  • Have you investigated remote display? (Score:3, Interesting)

    by Stinking Pig ( 45860 ) on Thursday November 22, 2001 @11:05PM (#2602444) Homepage
    In the server room, you'll be able to replace at least 70% if not 100% of the Windows computers.

    If you must replace the desktops, you'll have a tougher task. If you don't solicit feedback and get positive response from the users, I'd tread carefully.

    One way to ease the transition if you do whack the desktops is to provide a remote display service so that a couple of beefy Win2K boxes will supply the apps that people are currently used to -- that way you're not the bad guy, the support problem is centralized, and the performance is decreased enough that end-users will want to find Linux solutions.

    VNC is an option in this direction -- there's also apparently a Citrix ICA client for Linux.

    HTH,
    • I'd recommend investigating the remote display approach. Terminal server licenses are cheaper, and the management is a lot easier on a room full of servers, all of the similar spec, then a couple of thousand desktop machines. Also, you probably don't want to rule out:
      • people wanting to use Windows apps
      • people wanting to use Linux apps
      • people wanting to use a Windows desktop machine
      • people wanting to use a Linux desktop machine
      Tarantella [tarantella.com] can deploy windows and X apps to windows or linux clients. It can also use your directory server for authentication (if you have one), and has a powerful mechanism for determining which users get access to what apps.
      Linux/Staroffice cannot replace Windows/MSOffice completely, in any combination.
      [Disclaimer: I work for Tarantella]
  • i'd almost suggest splitting the migration into two parts. the first would concern servers, stuff the average user wouldn't necessarily see. the second would be the pc's in the labs that most of the students would end up using at one point or another.

    i'm sure there are plenty of guides to migrating over to servers, so i won't blab about it here, but for the lab migration i'd think about doing it one lab at a time, maybe taking the second most used lab and switching it, and then seeing what goes wrong/right with it. then use that as a kind of planner for how to migrate the other labs.

  • Don't do it. (Score:5, Informative)

    by jmaslak ( 39422 ) on Thursday November 22, 2001 @11:41PM (#2602556)
    I love Linux. I use it almost exclusively at home.

    But: I would not even consider what you are planning. You are considering computer costs, but not user costs. The costs of running the computing infrastructure (licensing, hardware, maintenance) are only some of the costs involved, and they are the small costs.

    The large costs are the cost of retraining 9000+ users. This is something you should not underestimate. Are those users going to be happy when they are running late, but can't figure out how to make Star Office (or whatever else) do what they know how to do in Word? Absolutely not.

    Some of the ENTERPRISE level problems with Linux currently:

    1) Think about directories. Any enterprise not implementing some sort of enterprise-wide directory needs to fire thier CTO. You need one source of information on all users that is stored in a central place and can be used by all applications. Sadly, MS is much closer to this then Linux right now. (Don't say "LDAP", either, since it is not supported in many applications - like kde/shell/whatever-else login!)

    2) Think about remote access. I've not been impressed by Linux's support for VPN. It's much better today then it was 3 or 4 years ago, but it is not done yet. If you use Windows, it comes out of the box (PPTP or, for the more security minded, L2TP). If you don't like either MS option, buy a third party option.

    3) Think about exchanging data. Ask your userbase how much data they exchange and with who. You might find that "PowerPoint Clone" isn't good enough. It doesn't matter why it isn't good enough - the fact is that people who exchange documents and require the document's formatting to be exchanged intact need to run the same program as the sender.

    4) Think about what your users know already. The less you have to change things, the better from thier standpoint.

    5) Think about databases. If you are really that large of a university, you will need some centrally administered databases. Databases which support huge datasets, stored procedures, transactions, foreign keys, etc. You might argue for PostgresSQL, but it won't stand up unless you find some reporting applications and such for the clients. Don't say Oracle, either, since Oracle on Linux is missing many features found in thier NT and Solaris offerings.

    6) Think about wierd hardware and integration with legacy machines. Right now, you do have that integration - in some way you can talk to all your machines. Make sure you don't break that. Think about people like EE and Physics, too, who might have some very wierd things hooked up to thier machines.

    Good luck - don't make the decision based on what either the Linux or MS lobby says. Instead, figure out where your enterprise needs to be 5 years from now and pick the software that supports the majority of the needs. Sadly, it may be MS software, since it does support directories, remote access, data exchange, existing user knowledge, databases, wierd hardware, and integration with other systems. Linux supports some of these, but does it support all of them?

    I haven't even mentioned things like PKI (not certificates, but actual infrastructure - things like automatic certificate renewal), wierd applications, etc, which I'm sure you'll figure out if you do a large scale study of where you want to be tomorrow.
    • 1) Think about directories.
      LDAP and PAM should be able to do the job
      2. remote access. freeswan, CIPE, ssh
      3. open standards for file formats should be ncouraged. PDF if nothing else.
      4. college is where they should learn, linux is great if you want to learn about computers, not so great if you just want to memorized one way of doing things.
      5) which features are missing? If you find linux/postgres inadequate stick with oracle on solaris. you can still dump NT. Stay away from SQL server.
      6)wierd hardware - case by case.

      linux is the platform of choice for open source software development.
      Don't go into this planning on replacing every MS box but you should be able to get rid of most of the servers painlessly. Just make sure you don't upgrade any desktops to later more restrictive licensing from MS. Take labs and individuals on a case by case basis. Setup a mirror and manage updates automatically. (autorpm, autoupdate, up2date) Mass deploy staroffice on linux/windows and solaris. upgrade to staroffice 6.0 when released.
    • Novell's eDirectory does pretty well under Linux, and it stomps all over Active Directory on NT/2000.
      It is, in fact, the best directory solution out there.
      Take a look at CNN's web-site sometime and see the little Novell logo in the bottom right-hand corner. eDirectory was the only thing which could handle their subscriber base.
      I've heard rumors that Yahoo's using it too, but can't confirm.
    • Most of your points are wise. Certainly, put the needs and abilities of the users first.

      I think you may be wrong about Oracle on Linux however. What features (other than client GUIs, which can be run from anywhere) aren't available on Linux? A counter-example: clustering [oracle.com].

      I suspect that even if they don't replace Windows in one stroke that they will find many opportunities to reduce costs with Linux.

      • I think he is probably referring to the GUI tools that come with the NT version of Oracle. Any DBA that is worth his/her salt is going to know how to use the sqlplus console.

        Doesn't Oracle develop their new implementation of their database products on linux first now?
        • Well, as I said, you can run the GUI tools from any machine, in fact preferably not on the server machine itself. So lack of GUI tools for Linux is no reason to avoid running the database server on Linux.

          I've done some development with Oracle on Linux and quite a bit on Solaris. Linux seems just as capable an environment for Oracle. I've also spoken to people who have experience scaling Oracle on Linux and have gotten good reviews.

          For much more, check out the oracle-on-linux mailing list [yahoo.com]. And here's a page at Oracle [oracle.com] that says Linux is supported "across the entire technology stack".

          But I don't want to plug Oracle on Linux too hard. You (for 95% of "you") probably don't really need to pay Ellison's sky high licensing fees. PostgreSQL [postgresql.org] really is quite good now and more than sufficient for most database applications, most of which barely scratch the surface of Oracle's scalability and capabilities.

          • For your information:
            Linux is now one of the Oracle strategic platforms (with solaris and NT). The support is at least as good as NT. Some tests have proved that performance is better with Linux. There are NO features missing in the last version of Oracle for Linux. It even includes all the GUI applications (Enterprise manager). I have had many good experiences with it. They are now bringing their dev. tools to Linux (JDeveloper is coming out soon)...
            I hope this clears it up
            Quentin
  • Can I've a job?

  • I don't have any experience in Linux installation of that scale, but I still have some suggestions:

    • Don't try a wholesale migration. Especially at a University, you've got a lot of different groups with a lot of different needs.
    • Go from easy to hard in order to build experience. The obvious first step is to replace the IIS boxes with Linux/Apache. It's easy and proven and there's really no reason not to do it.
    • E-mail and DNS could easily move next.
    • Next up, build a directory system using Linux.
    • Start building a nice Intranet so that people spend more time in their browsers and less in MS specific applications. This might be a good time to install Netscape or Mozilla everywhere so that the Linux transition is easier later on.
    • Then replace those file and print servers with Samba. By this time you should have a good deal of Linux experience to help you through the rough spots.
    • When StarOffice 6.0 ships, start loading it everywhere. Don't upgrade MS-Office, but don't push people to use SO yet either.
    • Finally, pick specific student labs for transition. Work with the instructors to make sure all the software the students need is available. Replace the labs one by one over time.
    • Lastly, the most difficult group is administration. They're stubborn and a lot less willing to learn new things. Again, it might help to go one department at a time. You may find that certain departments can't make the switch because of some legacy software.

    This kind of a phased approach should make things a lot easier and it'll be easier to sell to the administration. They can always back out or slow things down if there are problems. Also, if you phase it in over two years, by the time you get to your toughest customers - the admin staff - the Linux desktop interface and applications will have matured another generation.

  • We are almost finished implementing a change to our network. It is a WinNT network, with NT WS at all of the desktops, and ~50 remote users dialing in. We decided to move to Citrix Metaframe to run all of our apps (MS Office, Lotus Notes, and a 5250 emulator).

    For the desktops I put together a Linux build using Linux From Scratch, including X-Windows and the Metaframe client. At first I tried to get it running on a bootable CD, but had some issues with it. I built an NFS server which serves a root image for each of the types of system we have in the building (4 types x 2 video modes). Luckily, we don't have very many types of hardware, but there could be a set of scripts to determine hardware type before booting is complete. By using a bootable CD (using ISOLINUX), we can revert to the NT systems in an emergency (Metaframe on Win2000 can crash occasionaly). Also, since the server could crash, I put together an exact duplicate of the system and wrote scripts to switch to the other server in the case of a failure, by simply setting up an additional IP on the server's NIC.

    I also began to put together a firewall using Linux to replace our Firewall-1 on WinNT, but my boss chose to purchase a Firebox since it is easier to configure and set up. I found out that the Firebox is Linux-based anyhow, and I'm glad that we contributed to the cause (see Firebox's contributions to networking code).

    I also installed a Linux server over a year ago to handle internal web applications and for other odd applications, such as monitoring othe network devices and paging our cellphone in the case of a failure. Recently we also started using MySQL on the same server to provide address book functions to RightFAX users through MyODBC.

    My suggestion is to slowly replace individual servers or applications, not simply to put Linux in there, but to save money on each project. for instance, by switching to Linux on the workstations, we saved $75 each on Client Access Licenses, which are only needed if you are using Windows to access the domain.

    Someday we'll even switch to Linux/Samba for our domain controllers, probably instead of upgrading to XP.

Slashdot Top Deals

"If it ain't broke, don't fix it." - Bert Lantz

Close