LowEndBox - Cheap VPS, Hosting and Dedicated Server Deals

SecureDragon - $18/Year 128MB OpenVZ VPS in Jacksonville

SecureDragon Joe from SecureDragon has created some coupon code for the LowEndBox/LowEndTalk community, on their low end OpenVZ VPS plans.

With O128B you get

  • 128MB guaranteed/256MB burstable memory
  • 15GB storage
  • 200GB/month data transfer on 100Mbps
  • 1 vCPU core
  • OpenVZ/SolusVM

Promotion code has 20 uses each, or until 30 September, whichever comes first. Servers with GoRack in Jacksonville FL (test IP: SecureDragon has been around for almost a year now although KuJoe sounds a lot more experienced. Here’s a review from 96MB and there are some pretty positive feedbacks from their last promo here.

Latest posts by LEA (see all)


  1. Was expecting to see their XEN plans too

    September 20, 2011 @ 2:25 pm | Reply
    • By the way, I am using both their OpenVZ and XEN plans, very stable, highly recommended!

      September 20, 2011 @ 2:28 pm | Reply
      • Francisco:

        I suggest you benchmark again your OpenVZ VPS and compare the new results with the first ones.

        September 20, 2011 @ 4:22 pm | Reply
  2. 96mb? You mean that site of the guy who whines that his Nix Communication VPS has poor routing and somehow thats bad against the VPS he is “reviewing??

    September 20, 2011 @ 6:02 pm | Reply
    • Can you stop trolling everywhere?

      September 20, 2011 @ 8:36 pm | Reply
      • Sometimes people need to step up and point things out to others who do not see the truth or “trolling” as you call it

        September 20, 2011 @ 8:42 pm | Reply
        • Is the 96mb experience, and if he is wrong, is OUR decision to believe him or do further research.

          Anyway, is the only one here who owns a website with complete reviews about the main aspects of every provider, and has been very helpful to me. Period.

          September 20, 2011 @ 8:46 pm | Reply
        • There are actually others out there that are much better and feel free to actually read some reviews. Every single one is like THIS VPS’ SPEED SUCKS BECAUSE MY VPS IN CANADA GETS SLOW SPEEDS!!!

          September 20, 2011 @ 8:50 pm | Reply
        • @Joseph: You do not have to read about my reviews if you do not want to, and by the way, I have used 3 VPSes (only one of them in Canada and it was from Nix Communications :)) to test the upload speed.
          If you are a provider and you get a not so nice review from me, my recommendation is to see where the problems are and try to improve on it.
          And if you are just another consumer (just like me), then like what I said, feel free to forget about 96MB.com, you are more than welcome to since I was hoping 96MB would be a venue for some sensible discussions with mature people.

          September 20, 2011 @ 9:37 pm | Reply
    • And just for clarification, I have NEVER said Nix Communications have a bad routing, there are some problems with their networks, but in general, I think everyone on LEB (may be you are an exception) would agree with me that Martin provide pretty stable VPS and perhaps the cheapest in Canada, and that is why I used it for review and testing purposes.

      If you have any constructive suggestions on 96mb.com, feel free to let me know, my knowledge with VPS is definitely very limited and I am always looking forward for ways to make my reviews better. Thanks a lot!

      September 20, 2011 @ 9:41 pm | Reply
  3. Christian:

    Okay, I have to admit that securedragon is awesome. When they were first posted here on LEB I thought they would be the next deadpool provider, but 12$/year 96 MB VPS offer was so good I had to go for it. And they are rock solid, no downtime at all, the VPS itself performs well. I think securedragon is the second VPS provider(after BuyVM), who is able to provide great services at that ultra low budget. So, well done Kujoe!

    (but I hate the name “SecureDragon” I always forget their silly name and have to search them on LEB :D)

    September 20, 2011 @ 6:44 pm | Reply
    • The only downtime I had with them is approximately 30 minutes, according to Nagios, and that was scheduled maintenance.

      September 20, 2011 @ 8:21 pm | Reply
  4. Thanks for the kind words everyone. As for the 96MB review, I personally find their reviews to be pretty good and very detailed. If anything, their reviews have been extremely helpful for us and, as you can see from my comments to the reviews, we’ve made changes where it was needed and addressed what we couldn’t fix (i.e. Xen PV “bug” with network speed).

    I’ve found quite a few reviews written about us in the past and so far, knock on wood, they’ve all be mostly positive. I will make an attempt to round them all up and post them on our site along with links to them to give everybody a variety of reviews for those that need more information before hosting with us (and trust me, when it comes to hosting information is the key and you can never have to much of it before making a decision).

    As for the name, we have seen it misspelled a handful of time or referred to by one of our old brands which has been a bit difficult for us to fix but we hope that as we grow and our name becomes more common those problems will be a thing of the past. I originally picked the name Backup Dragon with no intention of ever offering anything other than backups but then I found LEB.com and the rest is history. :D

    Our priority is stability and uptime with a strong focus on support. We understand prices will always get lower and performance will always be higher so we try not to compete on those levels. We do plan on moving into the high performance market in a year or 2 (i.e. dedicated servers and 2-4 VPSs per server) but at the current time we are focusing on budget VPSs because this has always been where my head’s been at.

    As I’ve said many times in the past, my true passion is free shared hosting and I enjoy it immensely because I like providing a free service to people and seeing them turn it into something. While I cannot do this with VPSs, I feel that LEBs are a perfect balance for me and so I will continue to focus on them. Thank you all for your continued support and even those of you who are not clients of ours, thank you for playing an instrumental role in the LEB community.

    September 21, 2011 @ 6:47 am | Reply
  5. George:

    How about someone running the following performance test and posting results?

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

    September 21, 2011 @ 11:54 am | Reply
    • DD is not a “performance test” but here you go George: http://www.96mb.com/96mb-low-end-vps-review-part-xv-secure-dragon/

      September 21, 2011 @ 12:15 pm | Reply
      • Francisco:

        [DD is not a “performance test”]

        Could you please elaborate?

        September 21, 2011 @ 12:19 pm | Reply
        • Agreed with @KuJoe. Wikipedia says that “dd is a common Unix program whose primary purpose is the low-level copying and conversion of raw data.”

          September 21, 2011 @ 12:52 pm | Reply
        • Francisco:

          Funny. And what Wikipedia says about “performance test”?

          September 21, 2011 @ 12:54 pm | Reply
        • Actually we’re having this discussion on the forum.

          Here’s the thread: http://www.lowendtalk.com/discussion/214/detect-overselling-on-openvz

          Down towards the bottom.

          September 21, 2011 @ 1:03 pm | Reply
        • I can edit Wikipedia to say the dd command unleashes fairies and rainbows, which I’m trying to say that Wikipedia is not often a very accurate source of info.

          Try using ioping to figure out disk performance.

          September 21, 2011 @ 1:07 pm | Reply
        • Francisco:


          I´m talking about performance test not benchmark test.

          September 21, 2011 @ 1:07 pm | Reply
        • Right, for those who hate wikipedia, man dd says “dd – convert and copy a file” and “Copy a file, converting and formatting according to the operands.”

          September 21, 2011 @ 1:49 pm | Reply
        • Francisco:

          For those who love Wikipedia:

          Performance Testing covers a broad range of engineering or functional evaluations where a material, product, system, or person is not specified by detailed material or component specifications: rather, emphasis is on the final measurable performance characteristics. Testing can be a qualitative or quantitative procedure.

          Performance testing can refer to the assessment of the performance of a human examinee. For example, a behind-the-wheel driving test is a performance test of whether a person is able to perform the functions of a competent driver of an automobile.

          In the computer industry, software performance testing is used to determine the speed or effectiveness of a computer, network, software program or device. This process can involve quantitative tests done in a lab, such as measuring the response time or the number of MIPS (millions of instructions per second) at which a system functions. Qualitative attributes such as reliability, scalability and interoperability may also be evaluated. Performance testing is often done in conjunction with stress testing.

          September 21, 2011 @ 2:06 pm | Reply
  6. DD does not measure the performance of a system. It simply outputs the current write speeds at the time of running the command which is not an accurate benchmark of the system’s performance. While some people use it to determine disk I/O, it was not designed for that and thus there are much more accurate programs out there to determine that.

    September 21, 2011 @ 1:46 pm | Reply
    • Francisco:

      Sorry but you said [DD is not a “performance test”]

      I think you don’t know the difference between benchmarking a system and doing a performance test.

      September 21, 2011 @ 2:04 pm | Reply
      • Benchmark – an established point of reference against which computers or programs can be measured in tests comparing their performance, reliability, etc.

        September 21, 2011 @ 3:06 pm | Reply
        • But there isn’t any test that is really accurate.
          And dd IS a performance test IMHO.

          September 21, 2011 @ 3:53 pm | Reply
    • Its hard to understand what does the term “system” mean.
      Hence, its hard to accept that DD is noway related to “system performance” testing. I have done lots of comp. performance tests and first time I hear that.

      September 21, 2011 @ 4:04 pm | Reply
  7. SecureDragon is stellar. It’s the first host I’ve experienced that actually lives up to the often-made promise “no overselling!”. Performance and reliability is spot-on, and the company communicates extremely well.

    Kudos to KuJoe :)

    I’d love to get another one, but I’m forcing myself not to until I’ve shed some of the useless VPSs I currently have (from other providers).

    And I’m also hoping the KuJoe is going to make a 32mb package for LEB :)

    September 21, 2011 @ 8:16 pm | Reply
    • Well I did just make a 64MB plan last night for a custom quote and decided to keep it in the ordering system (just hidden), I’ll gladly make a 32MB plan and throw it in there also but it won’t be advertised on the main site right away. :)

      September 21, 2011 @ 9:35 pm | Reply
      • I’m interested :) Something like:

        32MB guaranteed memory
        64MB burst
        2GB Disk
        25GB Transfer
        1 CPU

        For $15 / annual?

        Note the 1 CPU instead of 0.5 :)

        Let me know if that’s do-able…

        September 22, 2011 @ 9:17 pm | Reply
        • That’s definitely do-able. The problem with this though is that the server would get overcrowded pretty fast if we sold such small plans. For every 1GB of RAM we could fit 32 32MB VPSs, after 8GB that’s 256 32MB VPSs, and at 10GB that’s 320 32MB VPSs on a single server. While I don’t expect such a thing to happen, we have to keep things like this in mind when we build our plans.

          That being said, I will put a limited amount of 32MB and 64MB VPS plans in WHMCS (won’t be advertised on our site) since I have no idea what kind of demand is out there for 32MB and 64MB OpenVZ plans. :)

          September 23, 2011 @ 6:29 am | Reply
        • @KuJoe: cool! Are they available yet? If so, how do I find them?

          September 23, 2011 @ 6:51 pm | Reply
        • George:

          Yes, how do we find these plans?

          September 25, 2011 @ 2:06 am | Reply
        • Just spoke with my partner about this and I will be adding some plans tonight. I’ll reply here when they’re available.

          September 25, 2011 @ 2:38 am | Reply
        • George:

          Would you happen to have a preview on the price/specs by chance? Interested, even if they’re not completely concrete atm.

          September 25, 2011 @ 2:41 am | Reply
        • Sorry for the delay, the 32MB and 64MB plans have been added.

          32MB RAM, 64MB Burst, 2GB Disk Space, 100GB Bandwidth, 1 IPv4 Address, and 1/2 CPU Core @ $13.90/Year
          64MB RAM, 128MB Burst, 3GB Disk Space, 100GB Bandwidth, 1 IPv4 Address, and 1/2 CPU Core @ $15.90/year

          October 2, 2011 @ 5:17 am | Reply
        • I was gonna get a 32MB for development but I’m not sure now. I’m seeing really erratic SQLite performance of the 96MB I currently have — http://blite.ca/?t=8

          Any other OpenVZ / SQLite users here?

          October 2, 2011 @ 1:27 pm | Reply
        • We’ve noticed a recent drop in disk I/O performance during the random DoS attacks. We hope that our new firewall, in addition to our OpenVZ Node Balance project, our VPS’s performance will return to normal. Unfortunately it will take about a week for our hardware firewall to arrive and be implemented so we’re stuck with the current random DoS attacks.

          We were trying to avoid controlling our user’s traffic but since most of them are unwilling to protect themselves we were forced to get involved. :(

          October 2, 2011 @ 11:37 pm | Reply
        • That’s encourage KuJoe. I’ll see how SQLite performance goes once you get that HW firewall in place, and I’ll probably get a 32MB’er then. But I see you’re offering just a 1/2 core….

          October 3, 2011 @ 9:28 pm | Reply
        • It didn’t make sense to offer a full core when a bigger plan (O96B) only got 1/2 a core. We’ve considered upgrading all plans to a 1 full core and utilizing the CPU priority options in OpenVZ but we are holding off until after we finish our balance project.

          October 4, 2011 @ 3:07 am | Reply
        • Random disk writes are really in the mud. Hope you get that firewall in soon.

          October 4, 2011 @ 8:58 pm | Reply
        • It arrived today and we’ll be setting up and configuring it tonight. It is scheduled to go into production Thursday morning after testing.

          October 4, 2011 @ 11:57 pm | Reply
  8. George:

    Signed-up today for a 128 box and so far a 5-STAR rating:

    1. Quick setup
    2. Nice compact Debian default install (so I didn’t have to uninstall a bunch of program I don’t use)
    3. The overloaded server indicator test as follows was quite satisfactory:

    dd if=/dev/zero of=test bs=16k count=64k conv=fdatasync;rm test -f; ./ioping
    65536+0 records in
    65536+0 records out
    1073741824 bytes (1.1 GB) copied, 21.334 s, 50.3 MB/s

    PS. The Apache2 install was a bit grumpy and did not want to start but running the following command and thereby changing how Apache does memory things solved the problem.

    aptitude install apache2-mpm-prefork

    September 21, 2011 @ 10:13 pm | Reply
  9. John:

    Just F.Y.I, These guys DONT allow IRC. It’s in their ToS

    September 21, 2011 @ 10:44 pm | Reply
    • George:

      Probably a small percentage of VPSers need IRC. :-)

      September 21, 2011 @ 11:01 pm | Reply
      • Spirit:

        … but mainly VPSers in low end box range :P

        September 21, 2011 @ 11:29 pm | Reply
    • Correct. Our data center does not like IRC and we do not like the trouble that has been associated with IRC (100% of all attacks against us since we opened have been IRC related).

      September 22, 2011 @ 4:16 am | Reply
  10. YDGH-Corey:

    As far as I’m concerned DD is an I/O performance test, even after reading all the comments I don’t know how it couldn’t be.

    September 22, 2011 @ 6:39 am | Reply
  11. Dan H:

    Well, I had one of the VPSs from this companies last offer.
    The support was great, the service was great, the uptime was fantastic…

    I’m not sure when this changed, ’cause I’m pretty sure it wasn’t in the TOS when I signed up, but they’re a no IRC stuff policy. So much for my bouncer server…

    September 22, 2011 @ 7:06 am | Reply
    • Unfortunately we’ve never allowed IRC because GoRACK had warned us about using it before we even installed our first server. I don’t want to fault GoRACK or make it sound like we pass the blame on them either, I’ve admitted in the past that I am a newbie to IRC and know very little about it which in itself was enough to add it to our TOS. Of course after recent events I realized we had good reason for disallowing IRC on our network.

      We’re really sorry about this and it was my intention from the beginning to find a second data center that can safely host nodes dedicated to IRC and VPN but so far we have had little luck in finding one that will fit our company.

      September 22, 2011 @ 7:56 am | Reply
      • Dan H:

        Oh well. I’m poking other hosts. So, yeah.. There’s another host that doesn’t say anything about IRC, their support hasn’t responded for 24 hours. This sucks.

        Hey, I don’t suppose its too late to get my znc config files.

        September 23, 2011 @ 6:24 am | Reply
        • Unfortunately we delete all files in violation of our Terms of Service after we suspend the VPS so that the client has the option of continuing to use our services without the offending files on the VPS when we return the VPS to the client. Sorry. :(

          September 24, 2011 @ 12:08 am | Reply
  12. Jim:

    You guys allow VPN? TUN/TAP enabled ?

    September 22, 2011 @ 8:25 am | Reply
  13. Out of the 20 uses for the O128B promotion 3 were marked as Fraud and never contacted us so I’ve added those 3 back in the wild, that means 3 uses left. ;)

    September 22, 2011 @ 8:37 am | Reply
  14. IoPsu:

    Any working and good discount code for the O128B plan (yearly payment)?

    September 22, 2011 @ 8:39 am | Reply
  15. IoPsu:

    KuJoe, the 16FORLEB code is not working, it never gives $18/Year….

    September 22, 2011 @ 8:49 am | Reply
  16. Both promotions are working. The 16FORLEB is for the O96B plan and the 18FORLEB is the O128B plan. ;)

    September 22, 2011 @ 9:03 am | Reply
    • Is a typo in the post, LEA posted 16FORLEB in both plans

      September 22, 2011 @ 9:59 pm | Reply
      • Good catch! The promo code was 18FORLEB for the O128B plan (which I believe is used up, not 100% sure until I log back in and check). We still have plenty of the 16FORLEB promotions left though.

        And yes, we do enable TUN/TAP on request (sorry I missed that comment earlier Jim).

        September 22, 2011 @ 11:37 pm | Reply
        • rkrazy:

          18FORLEB is all used up :[

          September 24, 2011 @ 12:35 am | Reply
        • Yup, that promotion went pretty quickly.

          September 24, 2011 @ 12:58 am | Reply
  17. David:

    KuJoe, do you own this? When I saw the thumbnail it reminded me of Host2x!

    September 24, 2011 @ 3:03 am | Reply
    • Yup yup! I took the red gradient colors from the old Host2x header (the one with the trees in it… boy do I miss that header).

      September 24, 2011 @ 3:41 am | Reply
  18. Just got off the phone with GoRACK with some great news!

    We should have IPv6 by the end of the week (right now only Level3 offers it but HE.net should be in there by next month to offer redundancy, not like Level3 ever goes down or anything though).

    We’ll also be obtaining our own ASN by next week and we are working with GoRACK to obtain the requirements to receiving our own IP space.

    Just wanted to let everyone know since IPv6 has been requested quite a bit from us.

    September 27, 2011 @ 11:38 pm | Reply
    • ab:

      please assign blocks instead of the practical joke that is “1x IPv6” by some comapnies

      thx in advance :)

      September 28, 2011 @ 1:17 am | Reply
      • I don’t care how many are assigned by default, as long as I can request (a reasonable amount) more for free. On OpenVZ, it’s stupid to assign a whole /64 as VPSs don’t have much control over networking.

        September 28, 2011 @ 2:57 am | Reply
      • What is wrong with 1 IPv6? Right now, 99% don’t use it, those that do, 1 is enough for most, those that need more, just ask.

        September 28, 2011 @ 3:44 am | Reply
        • snape:

          No kidding. There’s over a dozen posts wangsting about something most people aren’t even going to use…

          I figure somewhere out there is a hopeless nerd who wants “fault-tolerant redundancy!!!111” and is looking to run a half-dozen nameservers each on their own IPv6, and a dozen mailservers, each on their own IPv6, and a half-dozen Varnish caches, each on their own IPv6, with an Apache and a Lighttpd and an Nginx webserver backend, each on their own IPv6…

          …all on the same single $18/year VPS.

          That or they want to persistently troll some IPv6-enabled forum (/.?) through a VPN and hope the administrators only block single IP addresses…

          September 28, 2011 @ 12:21 pm | Reply
        • rm:

          1xIPv6, is to whoever understands that thing, just screams ‘clueless’, and will be a huuuge warning sign that a provider offering that may have similar pecularities in other aspects. So I’m glad that SecureDragon went with multiple IPv6 by default as they seem to be a quite nice company.

          Want some real world uses? An its own IP for each domain and subdomain hosted via HTTPS; without SNI tricks or expensive multi-domain certs. Sure, v6 not used much today, but do we provide a stuck-in-today service, or do we provide a future-ready service at no extra cost. Tough choice? :)

          September 28, 2011 @ 2:26 pm | Reply
      • Blackstorm72:

        Let’s not suffer from exhaustion of IPv6 by 2030 now because a few of us want a /4 block of them for their big VPS business lol.
        But even IPv6 has a limit no matter how huge it is compared to IPv4. It just needs functionality to the rest of the world really.

        September 28, 2011 @ 4:14 am | Reply
      • How about a /128 block? :P Seriously though, I completely agree with dmmcintye3 and Tim Flavin. We will probably assign 1 IPv6 address but allow users to request more.

        My biggest concern with IPs is abuse. In my years of hosting I have encountered more spammers and script kiddies than I’d ever wish on anybody. I can only imagine what kind of headache I would have on my hands if they all had hundreds of IPs at their disposal.

        September 28, 2011 @ 6:22 am | Reply
      • rm:

        > How about a /128 block? :P Seriously though, I completely agree with dmmcintye3 and Tim Flavin. We will probably assign 1 IPv6 address but allow users to request more.

        You have 16 zomgilions of IPv6, why not just give 16 addresses by default?
        Achieves two things:

        – no constant ridicule of your “1xIPv6” offer by IPv6 enthusiasts, and yes, there are ALWAYS remarks by many people whenever an offer with 1xIPv6 is mentioned on some IRC network’s #ipv6 channel or a v6-related forum – “Sheesh, WTF is this BS, they don’t get it, still old v4 thinking”, etc etc. Even 16xIPv6 causes raised eyebrows, but that at least has an excuse of “butbut solusvm can not doez subnetz and only manages IPs one by one”;
        – no support overhead from answering tickets and manually adding addresses.

        > My biggest concern with IPs is abuse. In my years of hosting I have encountered more spammers and script kiddies than I’d ever wish on anybody. I can only imagine what kind of headache I would have on my hands if they all had hundreds of IPs at their disposal.
        In the IPv6 world every customer gets at least a /64 or /56 subnet, and whoever is hassled by that person simply blocks the whole subnet.
        In the VPS environment, if you do not want to give a /64 per VPS (which some providers do), you can settle with something like a /120 (a contigious block of 256 addresses).

        September 28, 2011 @ 7:38 am | Reply
        • Anyone doing OVZ and offering a /64 is retarded.

          You have any idea how much even binding 256 addresses bogs down a VM’s boot time?

          While I would like to offer it as a block to make people like freenode happy, offering a customer even a /120 would make them complain about the spin up time.


          September 28, 2011 @ 7:50 am | Reply
        • rm:

          yep, by “some do that” I mean with Xen.

          September 28, 2011 @ 7:52 am | Reply
        • Thanks for the information. Maybe we’ll hold off on IPv6 for a while after all. :)

          September 28, 2011 @ 8:08 am | Reply
        • I’m curious as to why people would rather see IPs wasted than be used? I understand opening a ticket and requesting more IPs is tough for a lot of users but keep in mind that if we handed out 16 IPs per client we would be wasting thousands of IPs. I am aware of how many IPv6 addresses there are out there, but because there is an abundance of something it shouldn’t mean it’s acceptable to waste it.

          September 28, 2011 @ 8:22 am | Reply
        • rm:

          and also if a subnet is routed to a VPS it doesn’t mean it has to have every address in the subnet assigned to some interface, does it? That would be next to impossible with a /112 already. But you can just tell the user “you may use any address within your 2001:db8::/112”, and they can utilize however many addresses they want with “ip addr add…”. Routing-wise, this can be routed via some single IPv6 of theirs, and even OpenVZ should play fine w/that.

          Same question can be asked to you, why you prefer to waste addresses by leaving them unassigned, instead of providing them to your clients. After all that costs nothing to you, and allows them to immediately utilize multiple addresses if they so desire, rather than jumping through additional hoops posting tickets and I suppose even justifying the need for the IP addresses to you.

          September 28, 2011 @ 8:39 am | Reply
        • Not on OVZ alas. OVZ requires the admin to bind off or to give each server an eth device and have a custom system in place to apply the routes needed. Giving people a bridged interface without any sort of lock down is going to lead to abuse if you’re not careful.

          Even our portable IP’s system is going to have a fairly large amount of lock downs in place.


          September 28, 2011 @ 9:02 am | Reply
        • Spirit:

          wasted? People get real… only one who can waste those IPs are hosts, KuJoe as example by leaving them unassigned. It’s not like he would have once few billion customers to use all available addresses assigned to him.

          …we will in fact end up with the same situation we have now for about 4 bio

          This of course isn’t realistic unless you’re looking in distance intergallactic future with colonies in entire universe but then IPv6 won’t be the solution anyway. As I said and I will repeat again looking at IPv6 from old IPv4 perspective wont bring you nowhere. Get rid of that.

          Or like Aaron at pthree.org said:

          If there are exactly 6.5 billion people on the planet, and each person lived in their own house, and each of house had 6.5 billion light switches (I think we’ve grossly overdone an even remotely accurate representation, no?), we would still have left over 340,282,366,920,938,463,421,124,607,431,768,211,456 addresses. Yeah- we’re barely scratching the surface with that one.

          While this “one million IPv6 addresses” concern here serve us for little bickering and arguing to spend some time… it’s nothing more than that.

          Some more info (but I am pretty sure that most of you are able to find it by yourself of course)

          |||| |||| |||| |||| |||| |||| |||| ||||
          |||| |||| |||| |||| |||| |||| |||| |||128 Single end-points and loopback
          |||| |||| |||| |||| |||| |||| |||| ||124
          |||| |||| |||| |||| |||| |||| |||| |120
          |||| |||| |||| |||| |||| |||| |||| 116
          |||| |||| |||| |||| |||| |||| |||112
          |||| |||| |||| |||| |||| |||| ||108
          |||| |||| |||| |||| |||| |||| |104
          |||| |||| |||| |||| |||| |||| 100
          |||| |||| |||| |||| |||| |||96
          |||| |||| |||| |||| |||| ||92
          |||| |||| |||| |||| |||| |88
          |||| |||| |||| |||| |||| 84
          |||| |||| |||| |||| |||80
          |||| |||| |||| |||| ||76
          |||| |||| |||| |||| |72
          |||| |||| |||| |||| 68
          |||| |||| |||| |||64 Single End-user LAN (default prefix size for SLAAC)
          |||| |||| |||| ||60 Some (very limited) 6rd deployments
          |||| |||| |||| |56 Minimal end sites assignment[3] (e.g. Home network)
          |||| |||| |||| 52
          |||| |||| |||48 Typical assignment for larger sites
          |||| |||| ||44
          |||| |||| |40
          |||| |||| 36 possible future Local Internet registry extra-small allocations
          |||| |||32 Local Internet registry minimum allocations
          |||| ||28 Local Internet registry medium allocations
          |||| |24 Local Internet registry large allocations
          |||| 20 Local Internet registry extra large allocations
          ||12 Regional Internet Registry allocations from IANA[4]

          A /64 contains 2^64 nodes or 18,446,744,073,709,551,616 IPv6 addresses.

          That’s IPv6 subnet which most of us have either from vps provider either from tunnel broker and it’s all perfectly fine, isn’t it? So lets argue a bit more regarding those IPv6 addresses to spend some more of our free time…

          – RIRs assign ISPs at least a /32 prefix (enough for 65,500 business customers or at least 16 million home users);
          – ISPs assign their corporate customers with /48 subnets (giving each business 65,500 subnets to utilise throughout their organisation);
          – ISPs assign their home or small end users either /56 subnets (if under 256 subnets are required by the customer), or one /64 subnet (if the client does not require any further subnetting).
          – Each ISP, corporate client and end user will subnet their /48 or /56 prefix according to geographic region, department and end user, ensuring a true hierarchy of addressing and allowing efficient aggregation of address space throughout their organisation.
          – Each host will sit on a /64 subnet, no longer and no shorter.

          This subnetting policy is far more efficient than IPv4 and should keep the IPv6 routing table compact. It is anticipated that each ISP will have only one /32 prefix, making the global table small. In addition, each ISP’s internal routers are kept small since each of their corporate customers will have only a single /48 prefix. Home and small office users have the benefits of plenty of static IP addresses, eliminating problems often associated with dynamic IPs and use of Network Address Translation (NAT).

          I prefer to get some decent /IP block simply because:
          1. with own nameservers I can set up rDNS for those hosts by mysefl
          2. i wont ever again bother my hosts with request for new IPs

          September 28, 2011 @ 9:04 am | Reply
  19. It’s not wasting them by saving them to be used…. infact, it’s the opposite of waste.

    That being said, I think I have found a solution that will make all parties happy (if not, then the unhappy parties are more than welcome to start their own VPS company and manage it however they see fit). We will give the client the option to specify how many IPv6 addresses they need initially (unsure of the max number at this time). If they need more then they can open a ticket (which I guarantee does not have any hoops involved). ;)

    September 28, 2011 @ 8:59 am | Reply
    • Spirit:

      How can you justify so many (billions) IPv6 addresses which you wont ever put in use?

      September 28, 2011 @ 9:05 am | Reply
      • Only time will tell. :)

        September 28, 2011 @ 9:12 am | Reply
        • Spirit:


          September 28, 2011 @ 9:13 am | Reply
        • How so? We don’t even have IPv6 yet and you’re speculating on how we plan to use them?

          September 28, 2011 @ 9:15 am | Reply
        • Spirit:

          I do not speculating. I see only what you write and express my opinion regarding this, nothing more.
          btw. may I ask you what IPv6 range you expect to get? /48? /64?

          September 28, 2011 @ 9:21 am | Reply
        • No idea TBH, whatever GoRACK/ARIN will give us is what we get.

          September 28, 2011 @ 9:29 am | Reply
        • The smallest ARIN will give you is a /32, so 2^96 addresses :)

          GoRACK will give whatever it may be, probably a /64 or a /48


          September 28, 2011 @ 9:36 am | Reply
        • Spirit:

          Don’t be naive :) Resistance is futile. Soner or later your virtualization will support /range by default, sooner or later your control panel will support /range by default and sooner or later /range allocation will become standard which everyone, including you will offer :)

          This nowaday is just IPv6 tasting and most people are completely happy with few IPv6 addresses however with changing this will change also naive past objections and most of you will do the same as others – offer decent IPv6 range. We all like evorack and their preparation to add our own nameservers to given /64 range and that way give us control over our rDNS. Sooner or later this will become standard and you will not want to be behind.
          So it doesn’t matter what and how many IPv6 addresses you offer this moment. What it matters is that you will change your approach regarding IPv6 allocation and go against your nowaday aguments :)

          September 28, 2011 @ 9:48 am | Reply
        • Ville:

          @ Francsico edis.at gave me a range of 65000 IPv6 IPs. Don’t know what to do with them yet ;/

          September 28, 2011 @ 10:03 am | Reply
    • For what it’s worth there really is no spam over IPV6. The only abuse you really need to worry about is:

      – Some user going on efnet and getting hit by the rare IPV6 botnets. I doubt many DC’s know what to look for in an IPV6 DDOS.
      – Torrents, but contents owners aren’t really DMCAing over this since it’s pretty annoying to trace (nor their automated scripts support it as far as I know)

      #1 is the big one since we’ve already seen/had cases where users use an IPV6 connection to lock/steal nicknames from people. For the time being the big IRC botnets that do multiple gbit aren’t really IPV6 ready, but I expect that to change pretty quick.


      September 28, 2011 @ 9:05 am | Reply
      • Spirit:

        Irrelevant. You don’t need /96 to steal nicknames. Few IPs offered by you or KuJoe is more than enough for that. Beside that some of you making own “subneting policy” ingoring “subneting white paper” with making up stories like… IPv6 will be exhausted that way – which of course isn’t true.

        September 28, 2011 @ 9:12 am | Reply
        • Hense why I said that was one of the abuse things he has to worry about :)

          We’ve had cases of this on users with a 128MB and 16 ip’s.


          September 28, 2011 @ 9:21 am | Reply
      • Thanks for the heads up. IRC is already disallowed so when we find it on our servers we stop it before it becomes a problem. While I hate torrents, I don’t hate DMCAs so not a problem there. :)

        September 28, 2011 @ 9:14 am | Reply
      • Spirit:

        Btw. most hosts which assign /64 or so know very well which range every user have, usully they assign even customer number into their IP block as example customer nr. 102 have :0102::/96 prefix. I can’t see how this would be hardly to trace than customer with few random /128 IPs.

        September 28, 2011 @ 9:17 am | Reply
        • I actually like this idea, I will jot this down for later. Thanks!

          September 28, 2011 @ 9:20 am | Reply
      • ab:

        I had an old customer try to take the nickname ‘f’ on efnet.

        First time seeing 10gig dos over ipv6, mostly originating from native home connections in europe and a few dedis, lol

        September 28, 2011 @ 9:42 am | Reply
    • Spirit:

      With giving them some “small” IPv6 block, as example /96 or /112 you wont be ever again bothered with request for new IPs. That’s efficiency.

      September 28, 2011 @ 9:08 am | Reply
  20. Spirit:

    With old IPv4 approach to IPv6 some of you just makes things harder for yourself – long term :)

    September 28, 2011 @ 9:24 am | Reply
    • Worry not. I’m not a networking guy nor do I claim to be so when it comes to IPv6 my choice will not be the final decision without consulting with our DC network admins. I must say your posts have been very informative and, while I have no intention of picking up another networking book, I do need to rethink some of my previous statements and do some real planning before implementing anything (although I do like the idea of letting clients decide how many IPv6 addresses they want, the method you mentioned about about using their client ID for their IP block is nice also but might be hard to implement).

      September 28, 2011 @ 9:34 am | Reply
  21. john:

    What is going on, all servers are down slow time out.

    September 30, 2011 @ 3:54 am | Reply
    • My Xen box is up and running like it usually does and has not been rebooted since I last rebooted it.

      September 30, 2011 @ 4:05 am | Reply
  22. All servers are online but we’ve been experiencing quite a few attacks lately against our IP ranges with SSH attacks (bruteforce attacks that turn into DOS attacks). While the attack is not devastating, it is causing some issues for some VPSs and clients. We are working on a solution right now and apologize for the inconvenience this has caused you. I wish we could just block port 22 and be done with it but that’s not an option unfortunately. :(

    September 30, 2011 @ 4:37 am | Reply
  23. MIKE:


    June 6, 2020 @ 12:50 pm | Reply

Leave a Reply

Some notes on commenting on LowEndBox:

  • Do not use LowEndBox for support issues. Go to your hosting provider and issue a ticket there. Coming here saying "my VPS is down, what do I do?!" will only have your comments removed.
  • Akismet is used for spam detection. Some comments may be held temporarily for manual approval.
  • Use <pre>...</pre> to quote the output from your terminal/console, or consider using a pastebin service.

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