On Suricata 1.3, 1.4 and “next”

So with 1.4 out the door we have a new stable. However, we’re keeping 1.3 around for a few more months to give everyone the chance to plan updating to 1.4. Of course, we think 1.4 is a lot better than anything we released before, so we do recommend updating as soon as you can.

Continued support for 1.3 means we’ll do more releases to fix critical issues. We’ll probably include trivial fixes of smaller problems. When talking about critical issues I mean crash cases mostly. Anything else will be fixed only in 1.4 and up.

Moving forward, we’ll open our dev branch on 1.5 (or shall we jump to 2.0 already?) after the 1.4.1 or 1.4.2 release, which I hope to be doing sometime in mid-January. But we’ll see how things go.

If you have patches you need to get included, please open a pull request on github. Also, I think it’s a good idea to announce those requests on the oisf-devel list. So everyone on the list is notified and can help review and test.

Suricata 1.4 is out

About 5 months after 1.3 came out we’ve released 1.4, and we’ve been quite busy. Eric Leblond’s post here has all the stats and graphs. There are three big new features: unix socket, ip reputation and luajit. For each of these the same is true: it’s usesable now, but it’s the potential that we’re most excited about. Over the next months we’ll be extending each of those to be even more useful. We’re very much interested in ideas and feedback.

Performance obviously matters to many in the IDS world, and here too we have improved Suricata quite a bit again. We now have Suricata 1.4 running on a ISP 10gbit/s network on commodity hardware with a large ET ruleset. Of course, YMMV, but we’re definitely making a lot of progress here.

Sometimes the little things matter a lot as well. A minor new feature is that live “drop” stats are the the stats.log now:

capture.kernel_packets    | AFPacketem21              | 13640581
capture.kernel_drops      | AFPacketem21              | 442864
capture.kernel_packets    | AFPacketem22              | 7073228
capture.kernel_drops      | AFPacketem22              | 9449
capture.kernel_packets    | AFPacketem23              | 10528970
capture.kernel_drops      | AFPacketem23              | 148281
capture.kernel_packets    | AFPacketem24              | 7212584
capture.kernel_drops      | AFPacketem24              | 12643
capture.kernel_packets    | AFPacketem25              | 9763439
capture.kernel_drops      | AFPacketem25              | 17874
capture.kernel_packets    | AFPacketem26              | 10464106
capture.kernel_drops      | AFPacketem26              | 20378
capture.kernel_packets    | AFPacketem27              | 8869182
capture.kernel_drops      | AFPacketem27              | 18336
capture.kernel_packets    | AFPacketem28              | 7925045
capture.kernel_drops      | AFPacketem28              | 258168

This is supported for AF_PACKET, PF_RING and libpcap.

Last August we’ve added Suricata to github to make it easier to participate. Also, the code review tools associated with the pull requests are very useful. Github has been an unexpected success for us. At the time of writing there are 24 forks of Suricata on it, I’ve processed about 250 pull requests. The patches that have been submitted range from small fixes to full blown features, and more are on the way. I’m very grateful for these contributions and everyone’s patience with me.

Now that 1.4 is out, we’ll be taking it slow over the holidays. The team has been working like crazy, and everyone deserves a break. So the next weeks we’ll focus on further consolidation, fixing bugs that no doubt will pop up. Other than that, things will be slow. After the holidays we’ll start planning for the next milestone. Again, your ideas and contributions are very welcome! 🙂

First stable release in the 1.4 branch, right in time for the holidays. Enjoy!

Suricata

Photo by Eric LeblondThe OISF development team is proud to announce Suricata 1.4. This release is a major improvement over the previous releases with regard to performance, scalability and accuracy. Also, a number of great features have been added.

Get the new release here: suricata-1.4.tar.gz

The biggest new features of this release are the Unix Socket support, IP Reputation support and the addition of the Luajit keyword. Each of these new features are still in active development, and should be approached with some care.

The 1.4 release improves performance and scalability a lot. The IP Defrag engine was rewritten to scale better, various packet acquisition methods were improved and various parts of the detection engine were optimized further.

The configuration file has evolved but backward compatibility is provided. We thus encourage you to update your suricata configuration file. Upgrade guidance is provided here: Upgrading_Suricata_13_to_Suricata_14

New features

  • Unix socket mode for batched processing of…

View original post 584 more words

IPv6 Evasions, Scanners and the importance of staying current

Lots of activity on the IPv6 front lately. There was a talk on a conference on bypassing IDS using IPv6 tricks. Also a new scan tool (Topera) claimed to scan a host while staying below the radar of an IDS was released. To start with the latter, even though Suricata doesn’t have a dedicated port scan detector, the tool’s traffic lights up like a Christmas tree. The trick it pulls is to pack a lot of duplicate DST OPTS extension headers in the IPv6 packets. These options are just fillers, the only options they use are the “pad” option. In Suricata we’ve had an event for duplicate DST OPTS headers since 1.3 and the padding only headers generate an event in 1.4. Both alerts will be very noisy, so calling this a stealth attack rather dubious.

The other thing was a talk on IPv6 evasions, where the author compared Snort and Suricata. Suricata didn’t do very well. Sadly the authors chose not to contact us. On closer inspection it turned out an old Suricata version was used. Which one wasn’t specified, but as they did mention using Security Onion, I’m assuming 1.2. In the 1.3 branch (current stable) we’ve fixed and improved IPv6 in a lot of areas. Nonetheless, while testing the various protocol tricks, we did find some bugs that are now fixed in the git masters for the 1.3 stable branch and the 1.4 development branch.

I think these developments serve as a reminder that staying current with your IDS software’s version is critical. For that reason it’s too bad that distro’s like Security Onion, Debian, Ubuntu all lag significantly. The reasons differ through. For the guys from Security Onion it’s mostly a time problem (so go help them if you can!) for Debian and Ubuntu it’s actually policy. For that reason we’re providing PPAs for Ubuntu and for Debian we’re working on getting Suricata into the “backports” repo. The only mainstream distro that does it right for us is Fedora. They just update to the latest stable as soon as it’s out.

Given the complexity of protocols like IPv6 and the new developments all over the board, I see no viable case for staying on older versions. I know it’s a hassle, but stay current. It’s important.