Wednesday, September 26, 2007

Go Blue! Recruiting at Michigan (day 2)

Oh my am I exhausted! I hoped to have most of the text of this completed before my flight got back to Manchester last night, but that didn't happen.

I keep telling people I know that Michigan is a hardware school (in spite of having some great software people - see my post from Monday). We Solaris developers at the Sun table were brutally reminded of this yesterday. Lots of EE's with Verilog and/or VHDL experience. Many of them asking about architecture and/or verification, but a surprising number who have never heard of SPARC, the UltraSPARC T1 (aka. Niagara), or that they can see the entire source for the Niagara with OpenSPARC. Almost every business card of mine I handed out to folks had the word, "OpenSPARC" on the back so they could Google it later.

We also tried to make sure everyone had OpenSolaris disks. There are four binary distributions of OpenSolaris on that set of disks: Solaris Express Community Edition (see the previous link) - Sun's current OpenSolaris vehicle, Nexenta - which is probably going to be one of the more comfortable ones for Ubuntu Linux users to land in, Belenix - which is optimized for Live CD use, and Schillix, which was the first non-Sun distribution of OpenSolaris, by Joerg Schilling of "cdrecord" fame. I hope some of the students went home and had success playing with OpenSolaris. You all should visit and engage the community discussions with your feedback and questions.

I mentioned Monday about how much like a geezer I felt. I had more of that yesterday not only saying, "Class of '91" a few times, but also when Professor Quentin Stout visited our table. My only graduate-level class I took at U. of M. was his Parallel Algorithms class in the fall of 1990 (during Football/Marching Band season). Back in the day it was all theory - we discussed how to partition problems using the abstract PRAM (Parallel Random Access Machine). It was the ONLY parallel ANYTHING class offered when I had an available slot. This was when shared-memory multiprocessors were experiments or startups (anyone remember the BBN Butterfly, the Sequent Balance, or the Encore Multimax?). I mentioned to Prof. Stout I took his class back then. He then proceeded to tell me how the class is far more practical now. He told me all about stuff like OpenMP, and other high-level constructs that as a systems' programmer I just don't get to use all that much. I still, however, felt pretty smart for seeing the future back in 1990. I hope I have as good luck 17 years later.

Anyway, I had a great time in Ann Arbor, and I hope to get back there sooner rather than later. If anyone who visited our table is reading this, leave a comment, and don't be afraid to be honest. :)

Monday, September 24, 2007

Go Blue! Recruiting at Michigan (day 1)

I mentioned I was going to be at the University of Michigan's Engineering career fair, and here I am!

I got in yesterday (Sunday) afternoon, and did some things to re-orient myself. I visited my fraternity house first, and quickly, because rush began that night. In some ways things hadn't changed a bit - the house is still there and the rooms have the same names (my old room with a skylight window is still called Lighthouse). In other ways, they had - the TV is bigger and flatter, half of 'em had laptops, and the basement was being seriously renovated. The guys were pretty mellow, probably because of all of the post-beating-of-Penn-State celebrations. I then wandered around campus, eating dinner at Krazy Jim's Blimpyburger, where they give you burgers made of small, ground-that-day, patties. Yum!

When I flew in, the woman next to me on the plane explained the phenomenon she experienced when taking one of her kids to her alma mater. It all felt intimately familiar to her, even modulo some new buildings, but then she suddenly realized she was an old fart wandering campus. My kids aren't old enough to be shopping colleges yet, but I definitely felt the combination of familiarity and age. I saw buildings with new names, old names on new buildings, and just plain new buildings (esp. at North Campus). 20 years ago I was a freshman, now I'm literally old enough to be a father to a student in the incoming class of 2011.

This morning, I tagged along with Kais Belgaied as he visted some Computer Science faculty and grad students here. Our first visit was with Professor Z. Morley Mao, who's a new professor here. She has a lot of great ideas on how to exploit the Crossbow project for aiding intrusion detection (and mitigation), among other interesting ideas. We then talked to two other professors, Atul Prakash and Thomas Wenisch, and a few students as well. I remember Prof. Prakash from my time at Michigan (1987-1991), but the other two are new Assistant Professors. I'm confident from what I saw that U. of M.'s CSE division of EECS is going to be strong for a continuing number of years.

[Edit from Wednesday]Shoot! I forgot I also visited my old theory professor, Kevin Compton. He's a very good teacher, and helps even the most clueless undergrads (hem hem). He told me he's teaching a very popular undergraduate cryptography class, which is just too-cool, IMHO.

This evening several of us (Kais, Eric Kustarz, Bill and Sherry Moore, and I) gave a breezy tech talk about various goodies in OpenSolaris that we work on. We also had very yummy Pizza House pizza. Pizza House was "established 1986", which means it wasn't all that old when I was there, but it was good enough to have our host recommend it.

I'm now back in my hotel, squeezing packets over a flaky, but free, wifi. Tomorrow we will be spending the whole day at the table, taking resumes and answering questions. If one of you four readers of this blog is a U. of M. student, you don't have to wear a suit when visiting us. :)

Friday, September 21, 2007

More ZFS Love - Rapid Recovery

I recently scragged my laptop's primary root partition such that I needed to install-from-scratch again. I had a bootable secondary root, but since it was running an experimental BFUed build, this partition could not be upgraded.

Let's quickly look at how I configure my 100-decimal-GB laptop disk (&*%$% disk vendors):

  • c0d0s0 --> Primary root, approx 8GB (and I mean GB the way software geeks mean it, 8 * 1024^3).

  • c0d0s1 --> Secondary root, same size.

  • c0d0s3 --> swap, 3GB, same as main memory size (useful for system dumps).

  • c0d0s7 --> ZFS pool "tank", with 5 ZFS filesystems (tank, CSW, spro, local, and danmcd).

Before I shut it down for upgrade, I simply uttered this:

zpool export tank

That's it!

Then I plugged in my laptop to a local netinstall network, and PXE-booted to a Nevada build 73 install (which includes detangled NAT-Traversal) and started it up. I used the old Solaris installer because I know how to tell it to preserve disk slices. I told it to preserver the secondary root and the zpool.

One install later, I get root, and to recover my miscellaneous backups, CSW software, compilers, local binaires, and home directory, I just did:

zpool import tank

And again, that's it! All of my filesystems got mounted properly, no tables to edit, NOTHING.

I'll be at the University of Michigan Engineering Career Fair this coming Tuesday, and will be wandering campus on Monday. If you're one of the four people who read this blog and are there, drop by the Sun table - and see the very laptop I'm talking about. :)

Wednesday, September 12, 2007

IPsec Tunnel Reform, IP Instances, and other new-in-S10 goodies

Solaris 10 Update 4 (or as marketing calls it, Solaris 10 08/07) contains some backported goodies we've had in Nevada/OpenSolaris for a while.

IPsec Tunnel Reform was one of the first big pieces of code to be dropped into the S10u4 codebase. It shores up our interoperability story, so you can now start constructing VPNs that tell IKE to negotiation Tunnel-Mode (as opposed to IP-in-IP transport mode). Tunnels themselves are still network interfaces, but their IPsec configuration is now wholly in the purview of ipsecconf(1M). Modulo IKE (which we still OEM part of), we developed Tunnel Reform in the open with OpenSolaris.

Also new for S10u4 is IP Instances. Before u4, you could create non-global zones, but their network management (e.g. ifconfig(1M)) had to be done from the global zone. With u4, one can create a unique instance zone which gives the zone its own complete TCP/IP stack. The global zone needs to only assign a GLDv3-compatible interface to the zone (e.g. bge, nge, e1000g) to give it a unique IP Instance. You could have a single box be your router/firewall/NAT, your web server, and who knows what else, all while keeping those functions out of the fully-privileged global zone. It makes me think about upgrading to business-class Internet service at home, building my own box like Bart did and getting a few extra Ethernet ports.

Oh, and if you want to do it all with less ethernet ports, check out OpenSolaris's Crossbow and its VNIC abstraction!

Have fun moving your network bits in new and interesting ways!

Tuesday, September 4, 2007

Detangling IPsec NAT-Traversal, and a more stable API

As of OpenSolaris build 73, the way we do IPsec NAT-Traversal changes for the cleaner.

Before this build, IPsec NAT-Traversal was performed by pushing a STREAMS module on top of an open UDP socket. This module (nattymod) would either strip UDP headers out of ESP-in-UDP packets, or strip the "0-SPI" marker (four bytes of zeroes) before passing the datagram up to the application.

This method worked, but it had some flaws, including the implicit setting of certain socket options (UDP_INCLHDR) that would then potentially be blocked from applications that actually required them. Also, nattymod did not perform the insertion of the 0-SPI automatically, the application was stuck doing that on its own. And while FireEngine merged TCP in to IP for S10, we needed to wait for one of the earlier builds of OpenSolaris to get the UDP equivalent.

With the new NAT-Traversal scheme, a key management application (like our closed-source in.iked(1M)) that wishes to aid in NAT-Traversal simply sets a new socket option: UDP_NAT_T_ENDPOINT. If this option is set, the following things happen:

  • On inbound packets, the first four bytes after the UDP header are inspected.

    • If there are less than four byte, the packet is dropped, and assumed to be a NAT-T keepalive.

    • If the four-byte are all zeros (i.e. the 0-SPI), they are stripped and regular UDP processing occurs.

    • Otherwise, the UDP header is stripped and the packet is shuffled off the IPsec's ESP for processing.

  • On outbound packets

    • The 0-SPI is inserted between the UDP header and the user-generated data.

    • ESP will send ESP-in-UDP by itself depending on the Security Association's properties.

This will help anyone who wants to port their open-source IKE or other key management application to Solaris deal with the possibility of NAT boxes.

And on a related note, this will be mentioned during Nicolas Droux's OpenSolaris Networking for Developers talk next week at Sun Tech Days in Boston. I'll be there too, talking about S10 and OpenSolaris security features, as well as being in the audience for Nicolas's talk.