Forgot your password?
typodupeerror

DIY Carrier Grade Linux with Debian 75

Posted by timothy
from the cool-for-your-home-switchboard dept.
An anonymous reader writes "Carrier Grade Linux, once the domain of big-bucks Bells and commercial software vendors, just became more attainable for universities, companies running high-availability web services, and average Linux hackers interested in learning what goes into the world's most reliable, maintainable, and available systems. The Debian project, backed by HP, has launched the Debian-Carrier Grade Linux subproject, and registered Debian-CGL with version 2.02 of the CGL spec. LinuxDevices has created a simplified version of the registration form that lets you see which Debian packages to apt-get, and which packages you'll have to download and compile out side of Debian, in order to get your own Carrier Grade Linux setup."
This discussion has been archived. No new comments can be posted.

DIY Carrier Grade Linux with Debian

Comments Filter:
  • by spud603 (832173) on Monday June 05, 2006 @02:06PM (#15473947)
    An excellent example of the clarifying power of hyphens.
  • by Anonymous Coward on Monday June 05, 2006 @02:12PM (#15474003)
    Personally I'd rather wait for torpedo grade Linux.
  • by Nate Fox (1271) on Monday June 05, 2006 @02:14PM (#15474015)
    http://en.wikipedia.org/wiki/Carrier_Grade_Linux [wikipedia.org]

    Carrier Grade Linux' is a set of specifications which detail standards of availability, scalability, manageability, and service response characteristics which must be met in order for Linux to be considered "carrier-grade" (i.e. ready for use within the telecommunications industry). The term is particularly applicable as telecom converges technically with data networks and commercial off-the-shelf commoditized components such as blade servers.
  • by vishbar (862440) on Monday June 05, 2006 @02:17PM (#15474037)
    "Carrier-grade" means that the server pipes all incoming data directly to the NSA.
  • Wikipedia (Score:5, Informative)

    by flood6 (852877) on Monday June 05, 2006 @02:17PM (#15474039) Homepage Journal
    I was lost as hell over this summary and even TFAs. Here's some help [wikipedia.org], apparently "Carrier-Grade" refers to telecommunications carriers, which can typically accept no more than 30 seconds to 5 minutes of downtime per year from their servers.
    • Re:Wikipedia (Score:4, Insightful)

      by smbarbour (893880) on Monday June 05, 2006 @02:31PM (#15474148)
      Hmmm... Let me know when someone finds a "carrier-grade" carrier. I have yet to find any carrier with 5 minutes or less of downtime per year. Our current carrier is at approximately 24 hours of downtime per quarter-year.
      • Re:Wikipedia (Score:3, Informative)

        by parc (25467)
        You're mistaking carrier-grade voice service with a data SLA.
        Carrier-grade services require that a telco provide dialtone, routing, completion, and ring within certain time limits a certain percentage of the time. In all my years of dealing with telcos, I've only missed a dialtone twice. I've gotten failed routing far more often, but still well within 99.999% of the time. My data circuits, on the other hand, have failed often.
        • At least one of our PRI lines will go down for a day or two once a year. Maybe it doesn't count for the "carrier-grade" services when it's a problem with the physical media. SBC generally has to replace the cable run once per year (although I'm not sure why).
          • You should check your contract. Three nines of service is nine hours of (unscheduled) downtime per year. The vast majority of contracts will specify an SLA of four or five nines and you're barely getting three.
          • It all depends on what your SLA is with the carrier. Having worked for a wireless telco, I can assure you that outages are not taken lightly and that availability is a very high priority. Outages for a customer when a single line get cut or even for several if a minor trunk fails are not that uncommon, it's a crazy world we live in and few people are willing to pay for multiple connection points to different hubs. But unplanned outages of any hardware or software element at a switching station or hub are
      • Re:Wikipedia (Score:3, Insightful)

        by rcw-work (30090)
        Hmmm... Let me know when someone finds a "carrier-grade" carrier. I have yet to find any carrier with 5 minutes or less of downtime per year.

        Telcos feel they need 99.999% uptime from their equipment in order to provide you with a much lower level of service - typically 99.9% for a T1 or an analog voice line, occasionally 99.99% for a set of redundant circuits.

        Our current carrier is at approximately 24 hours of downtime per quarter-year.

        That's roughly 99%. If this is a T1, you should be able to do ten times

      • Telecom carrier? (Score:3, Interesting)

        by PhYrE2k2 (806396)
        Telecom carrier? I'd doubt it. Internet service provider- fine- but they're not a carrier.

        Telecom carriers are Long Distance providers, and Ma-Bell providers around the globe. They are the ones that provide power into your home for your phone service as well as the service itself. They are the ones that do switching entirely on closed circuts.

        Carrier grade is usually coined as 5-9's (9.9999%) which is friggin amazing. It's what the systems are designed for, and they usually pull it off.

        -M
        • Carrier grade is usually coined as 5-9's (9.9999%) which is friggin amazing. It's what the systems are designed for, and they usually pull it off.

          Yeah. Amazing they're still in business.

        • The problems we have are always in the SBC-owned lines. Oddly enough though, it's never on the T1 going to our connection to the backbone (Technically, we are our own ISP). It's always on either the voice PRIs or the point-to-point T1s. The problem is always a physical problem in the lines between our building and the origin of the lines (which is at the Motorola HQ a couple blocks away) Sometimes they just switch some pairs around. Other times it requires a new cable run.
    • Re:Wikipedia (Score:3, Insightful)

      by jc42 (318812)
      [A]pparently "Carrier-Grade" refers to telecommunications carriers, which can typically accept no more than 30 seconds to 5 minutes of downtime per year ...

      Well, if we measure this in a way comparable to the way that phone companies measure uptime, it'll mean measuring the time that the OS responds to pings. A machine that is a total zombie, with no processes making any progress, will be considered "up" if you can ping it from a nearby machine.

      After all, we are all familiar with phone systems that give a d
      • Carrier grade means not only uptime of the machine, but also that the machine correctly responds to requests within the limits set out in a Service Level Agreement (SLA).

        But note that telcos usually only offer service with an SLA to customers paying through the nose.
      • Re:Wikipedia (Score:3, Interesting)

        by Khyber (864651)
        The on-screen clock changes once a minute.

        Could that be simply because OSX and Mac OS (versions 9 and below,) do *NOT* show the seconds that pass by? *points to his default-install 72-D version iBook running OSX 10.2.8 and his Powerbook 190c running Mac-OS 8*
  • Debian Is Top Dog (Score:4, Insightful)

    by devphaeton (695736) on Monday June 05, 2006 @02:19PM (#15474059)
    Debian has long been 'the example' IMHO. RedHat got all the fame and glory, but Slackware and Debian really showed what Linux should be like.

    I just wish all these projects (i.e. ubuntu) that base off of debian would give them more credit.
    • by JerkBoB (7130) on Monday June 05, 2006 @02:34PM (#15474166)
      I just wish all these projects (i.e. ubuntu) that base off of debian would give them more credit.

      How?

      From About Ubuntu [ubuntu.com]:

      Ubuntu is a free, open source operating system that starts with the breadth of Debian and adds regular releases (every six months), a clear focus on the user and usability (it should "Just Work", TM) and a commitment to security updates with 18 months of support for every release.


      It's right there in the first sentence... Perhaps you want a large blinking banner at the top of ubuntu.com?

      A large number of the Ubuntu devs are (wait for it...) Debian devs, too. Ubuntu regularly contributes back to Debian. I'm sure there are political squabbles, but to say that Ubuntu doesn't give credit to Debian is nonsense.

      Bleh.
      • Ok... Ubuntu was a bad example on my part.

        There were a number of distributions about 4 years ago, many of which have 'folded' by now that would initially claim "FooLinux based on Debian".

        Ok great.

        But then they'd go on to say "FooLinux's innovative package system called Foomaptic-get" which was a symlink to apt-get or aptitude, and it was pointed at FooLinux's mirror of Debian unstable, everything renamed. Stuff like that.

        So I don't have other examples or sources, which makes my point appear weak and theref
    • Ubuntu isn't secretive about it all... anyone who knows even close to enough to do anything with the knowledge that Ubuntu et al are based on debian (and that includes financial donations as well as coding efforts/support/whatever) is more than well aware of the relationship.
    • by LWATCDR (28044)
      "Debian has long been 'the example' IMHO. RedHat got all the fame and glory, but Slackware and Debian really showed what Linux should be like."
      That is your opinion. There is no one size fits all.
      Some people want a simple, easy to install Linux.
      Some people want the latest and greatest Linux
      Some people want the most stable Linux that will provide the most up time.
      There are many projects that show what Linux should be like. Some projects are useless or redundant but if that makes there developers happy so be i
    • Debian has long been 'the example' IMHO. RedHat got all the fame and glory, but Slackware and Debian really showed what Linux should be like.

      Slackware showed us what, that Linux shouldn't have package management? And Debian has showed us what, that packages should be old and outdated?

      I've moved to Ubuntu, like most of the rest of the world, but it doesn't make me believe that Debian is the ultimate Linux, or that it does everything better than everyone else.

      Formerly I used Gentoo. Maybe you only

      • And Debian has showed us what, that packages should be old and outdated?

        This is not a drawback. Rethink this. It has a stable branch for a reason; if you don't like that aspect, then testing is for you.
      • 1) Slackware does have package management. It's called pkg_add.

        2) Debian STABLE is old as dirt. Debian testing or unstable is the cutting edge. But iff you want zero-day software, you can always compile it yourself from source.

        3) About the video resolution- If it won't auto-detect what you want, can you not just adjust /etc/X11/Xorg.conf?

        And yes, I understand that not every linux is for everyone.
        • 1) Slackware package management is a sad joke.
          2) Stable doesn't have to mean ancient. Debian is well behind the curve of stable released versions. They're behind because they're testing their customizations. IMO they need either more package maintainers (I know, hard to come by) or less customization. Then again, this is why I don't use Debian, and I have that choice - but this is a seriously common complaint.
          3) Actually, I've been trying to do this, but the configuration is somewhat strange (mode clocks
          • 2)

            I like 'unstable' distros - the latest and greatest, thats why I use ~x86. But on a server, I'm glad that I can use mysql4.1, php4.3, etc without having to upgrade, play around with updated config file formats, workaround new bugs, etc - thats why I use debian sarge on the webserver at work (I run FreeBSD 6.1 on my home server). The right tool for the right job.
      • Oh and btw...

        Even though I recognize your nick, your comments make it sound like Ubuntu is the new $fanboy_linux_distro. For awhile it was Mandrake, then Gentoo, then Knoppix, now Ubuntu?

        I'm having a hard time keeping up, sorry. ;-)
        • Ubuntu is DEFINITELY the new fanboy distribution. BTW, it was Redhat -> Mandrake -> Gentoo -> Knoppix -> Fedora -> Ubuntu... You left out the early part of the chain. Note that four of five of these got their popularity due to ease of installation and use... Fedora had the shortest honeymoon, in between the time when it came out and had the most eye candy, and the time when redhat started permanently destroying people's hardware.
          • I knew RH was in there, but it's ancient history ;)

            I've been running FreeBSD for the last three or four years so i haven't stayed up on all this :D

            Oh, and don't forget the RedHat/Mandrake-->OSX Fanboy jump :D
      • And Debian has showed us what, that packages should be old and outdated?

        Debian, as we covered previously, tends to have the oldest, most outdated packages around. There may be reasons for that but it does users little to no good.

        You are just showing your ignorance. Debian Stable has older packages than Ubuntu but they're also better tested and less buggy. Then there are Debian Unstable and Testing that both have newer packages than Ubuntu.

        And both Debian Stable and Testing get security updates for ALL th

  • by Anonymous Coward
    apt-get install interceptors
  • I am a huge Babylon 5 Fan, and I always wanted to build a "Destroyer Class" Linux box meant to secure my networks from attack by hostile nodes.
  • by ortholattice (175065) on Monday June 05, 2006 @03:14PM (#15474515)
    Although the concept struck me as amusing, given that carrier-grade requires 99.999 (5 nines) to 99.9999 (6 nines) percent reliability, still I couldn't imagine that MS would allow itself to be trumped by this.

    And, sure enough: from Google, "carrier grade Linux" - 114000 hits, "carrier grade Windows" - 17 hits (but still, not 0). The top Windows hit is from 1998: "a Microsoft white paper available at SUPERCOMM '98 will discuss carrier-grade Windows NT Server-based systems." Well, at least they talked about it, you gotta give them credit for that. Haven't heard much about it since, though.

    • I wonder what the reliability is like on NT/Embedded these days. I bet it's pretty damned hot. NT's Kernel has (or at least had) a lot going for it, not least a team of geniuses who put it together in the first place.
    • Although the concept struck me as amusing, given that carrier-grade requires 99.999 (5 nines) to 99.9999 (6 nines) percent reliability, still I couldn't imagine that MS would allow itself to be trumped by this.

      Show me standard pc hardware that is five 9s before you start worrying about Microsoft.

  • So... does this spell the death ofNO CARRIER
  • Do we need this ? (Score:2, Insightful)

    by amias (105819)

    Seems like a good idea at first but if you have 5-30 minutes downtime per year
    that means one very quick kernel patch per year . If you are really concerned
    about uptime applying patches in a timely fashion is just as important as
    hardening the system to start with.

    Obviously starting with solid proven code should mean less patches are needed
    but nobody is perfect and what about new functionality ?

    That kind of uptime is IMHO more a function of your hosting environment and the
    hardware you choose , this is going t

    • That's why you have redundancy and spares, so that you can take a system down for preventative maintenance, hardware maintenance and software updates, without interrupting service to the customer.
    • But how about being able to patch an application whiles it's running or being able to field debug an application without degrading performance while its running in production? That's why we need a CGE of Linux. Monte Vista Linux has had a CGE of Linux out for a while now but it's not a free distro. It will be interesting to see if Debian can provide the same features that Monte Vista CGE provides expect at no cost.
    • Wouldn't virtualisation alow for this, for instance moving a working system with zen, alow for this to happen?

      Genuinely curious ...
  • I can think of no reasonble purpose for me to have this, yet, I want it. Bad. "Sure that's a nice 64 way system, but dude, it's not even carrier grade." ffoiii
  • And just what does CG cost the user? Just extra time and effort on his/her part, or out-of-pocket money for some of the modules and/or instructions?
  • by BrookHarty (9119) on Monday June 05, 2006 @04:52PM (#15475326) Homepage Journal
    Thats the dirty little secret, scheduled downtime. As long as you schedule the downtime, its still carrier grade. I've yet to see a service even with maintenance windows stay up for a month. Service in terms of big pile of servers running multiple applications with a big fat database cluster behind it. YMMV.
  • With the exception of an OS we will not mention 'carrier grade' to me at least means top notch redundant and hotswappable hardware (all hardware). I am sure some of the self healing stuff in solaris helps it with being carrier grade however. As far as linux goes, I think it is already 'carrier grade' based on the equipment and demand. I have a fedora 3 server (coldfusion, mysql, apache2, cgi-perl stuff, sunray server) that has been going non-stop for ages and another fedora 4 with less traffic that has
    • To some extent Linux is fault tolerant. Years ago I was running my firewall on an old Thinkpad 750Cs (Yeah... 25 MHz 486 ;) This was with a series 2.2 kernel. As I was preparing to leave for vacation it started making loud whining noises. And it does not have a fan. :( I warned my son tha the was likely to lose Internet connectivity and I would deal with it when I returned. Upon my return, I noted that the Thinkpad had gone silent. When I asked when it failed I got "It's still working." I logged in and foun
      • Amazing, I think the true test is always going to be on top notch equipment, built for the purpose. Since HP is dumping its own OS and moving to linux I find it not at all surprising that they are now trying to, perhaps, fud there way into some kind of ultra stable linux niche. Sadly all they really need to do is supply top notch support and they will be 99% of the way there.
  • Wow,

    Finally, A linux subject that the slashdot crowd is silent about. Since it references debian the ubuntu comments were inevitable.

    In reality this goes far beyond a debian discussion and is actually great news for the business of linux. In general Carrier Grade OS's is a way of saying that the OS used for certain carrier/telecom applications must follow a certain set of rules and standards.

    This is important because carrier grade linux has to support a ridiculous feature set in order to achieve guarantee

The Universe is populated by stable things. -- Richard Dawkins

Working...