Multipath TCP

During the IETF 75 meeting in Stockholm, there was this multipath tcp BOF (“start-up meeting” sort of) on Thursday morning that I visited.

Multipath TCP (shortened to MPTCP at times) is basically an idea to make everything look like TCP for both end points, but allow for additional TCP paths to get added and allow packets to get routed over any of the added flows to overcome congestion and to select the routes where it flows “best”. The socket API would remain unmodified in both ends. The individual TCP paths would all look and work like regular TCP streams for the rest of the network. It is basically a way to introduce these new fancy features without breaking compatibility. Of course a big point of that is to keep functionality over NATs or other middle-boxes. (See full description.)

The guys holding the BOF had already presented a fairly detailed draft how it could be designed both one-ended and with multiple adresses,  but could also boast with an already written implementation that was even demoed live in front of the audience.

The term ‘path’ is basically used for a pair of address+port sets. I would personally rather call it “flow” or “stream” or something, as we cannot really control that the paths are separate from each other as those are entirely in the hands of those who route the IP packets to the destination.

They stressed that their goals here included:

  • perform no worse than TCP would on the best of the single TCP paths
  • be no harder on the network than a single TCP flow would be, not even for single bottlenecks (network and bottleneck fairness)
  • allow resource pooling over multiple TCP paths

A perfect use-case for this is hosts with multiple interfaces. Like a mobile phone with 3G and wifi, as it could have a single TCP connection using paths over both interfaces, and it could even change paths along the way when you move to handover to new wifi access-points or when you plug in your Ethernet cable or whatever. Kind of like a solution to the mobile ip concept with roaming that was never made to actually work in the past.

The multipath tcp mailinglist is already quite active, and it didn’t take long until possible flaws in the backwards compatibility have been discovered and are being discussed. Like if you use TCP to verify that a particular link is alive, MPTCP may in fact break that as the proposal is currently written.

What struck me as an interesting side-effect of this concept, is that if implemented it will separate packets from the same original stream further from each other and possibly make snooping on plain-text TCP traffic harder. Like in the case where you monitor traffic going through a router or similar.

I host

Sara Golemon, the founder and former maintainer of libssh2, pointed over the main site to my server the other day and now my previously unofficial libssh2 web site suddenly turned out to be the only and official one.

The plan is now to get the web contents push into a separate git repo to allow all libssh2’ers to modify it.

I’m also open and interested in feedback and ideas on how to improve the web site in whatever kind of way you think. Consider the current site mostly a placeholder for the info we have. How can we make it better?


HTTP cookies IETF working group

So finally (remember I mentioned this list when it was created back in January 2009) an IETF http-state working group was created, with the following description:

The HTTP State Management Mechanism (Cookies) was original created by Netscape Communications in their Netscape cookie specification, from which a formal specification followed (RFC 2109, RFC 2965). Due to years of implementation and extension, several ambiguities have become evident, impairing interoperability and the ability to easily implement and use HTTP State Management Mechanism.

I’m on the list from the start and I hope to be able to contribute some of my cookie experiences and knowledge to aid the document to actually end up with something useful. The ambition, while it was “toned down” somewhat since the initial posts of the mailing lists, is still fairly high I would claim:

The working group will refine RFC2965 to:

  • Incorporate errata and updates
  • Clarify conformance requirements
  • Remove known ambiguities where they affect interoperability
  • Clarify existing methods of extensibility
  • Remove or deprecate those features that are not widely implemented and also unduly affect interoperability
  • Add features that are already widely implemented or have a critical mass of support
  • Where necessary, add implementation advice
  • Document the security properties of HTTP State Management Mechanism and its associated mechanisms for common applications

In doing so, it should consider:

  • Implementer experience
  • Demonstrated use of HTTP State Management Mechanism
  • Impact on existing implementations and deployments
  • Ability to achieve broad implementation.
  • Ability to address broader use cases than may be contemplated by the original authors.

The Working Group’s specification deliverables are:

  • A document that is suitable to supersede RFC 2965
  • A document cataloging the security properties of HTTP State Management Mechanism

I think this is a scope that is manageable enough to actually have a chance to succeed and its planning is quite similar to that of the IETF httpbis group. Still, RFC2965 lists a huge pile of stuff that has never been implemented by anyone and even though it was a while since I did read that spec I also expect it to lack several things existing cookie parsers and senders already use. The notorious IE httpOnly is an example I can think of right now.

My HTC Magic Review

This is my first “smartphone” I’ve owned myself so of course I have nothing else this fancy to actually compare against. I’ve played around with others’ a few times but that doesn’t really count. I’ve owned perhaps 8 mobile phones since I got my first one 1996, and they have all been Nokias and Sony Ericssons.

I was never really interested in iPhone due to many reasons. It is not open. It has a (very) restricted app distribution mechanism. It forbids apps from running simultaneously etc. And it has a pretty strong connection with itunes with no proper mass-storage syncing supported. But I admit that it has a slick UI and many cool apps.

My plan is to get some Android hacking going eventually and this is basically the first Android phone that has reached Swedish soil. I mean without requiring me to bend over backwards to get it, as I’m sure I could’ve bought previous Android phones from obroad if I really wanted to.

Random good things:

  • it’s fast, most things run faster than on my previous Sony Ericsson thing and yet this is way more advanced with much bigger screen estate and fancier UI
  • it has a nice gui that you mostly can guess how to work with
  • I love being able to use a qwerty-style keyboard when messaging instead of relying on T9 etc
  • wifi is fun, but with a decent data plan it basically only brings me slightly improved speed and I often can’t even tell the difference!
  • the integration with the Google services are nice, gmail and maps most noticeably
  • there really are a bunch of existing cool apps (I know iphone has lots more, but there are still thousands)
  • it has a much better approach to messaging, similar to what I’ve seen in the iphone, than I’ve ever experienced in a Nokia or Sony Ericsson. It focuses on conversations and keeps the “thread”.HTC Magic
  • I really really like the feeling of it being a networked thing that also can make phone calls. I can browse, use maps, use gmail just as easily as I can message or call people. With my previous phones all the internet-related services always felt tacked on like a very late afterthought.
  • The notification system is nice, and the three-screen wide “home” with its widget-system is really neat.

Bad stuff:

  • I’ve had some apps crash on me on occasion. But it’s rarely a problem as they’re restarted automatically for me.
  • Toggling wifi on/off a lot can sometimes lead to me not getting any data network at all, and I’ve had to reboot the phone to get back to phone-based (Edge/3G) data.

On-screen keyboard

Of course any and all geek friend I have ask me about how I deal with the on-screen keyboard. I must admit I’m still quite fond of it. Mostly because a physical keyboard makes the phone clonky and it adds physical contraints and wear-points that I don’t like. So the keyboard is a bit small, especially when the phone is in portrait mode, but the suggested completions are fine and I believe I’m already typing pretty quickly on the thing. When I ssh’ed from the phone to one of my servers I did find the obvious lack of cursor keys (to for example navigate an ordinary ncurses-based app or the command line history of a bash prompt) but other than that I really can’t complain.

Background Applications

One obvious advantage compared to iphones is of course the ability to run applications exactly the way I’d like. I can actually run the irc client and then have it in the background while I go browse the web or answer a call or whatever and then at my choice go back to the still connected irc client. In fact when playing with this it feels like a really ridiculous restriction of the iphone.

Comparing to my SE w550i

My previous phone is 94 grams compared to the Magic’s 116. The magic has a much bigger screen. The magic is roughly 11mm wider and 14mm taller. That makes it use 30% more volume (85 cm2) but still fits fine in the front pocket of any set of pants I use. The magic claims a lot longer battery life, but given that it has so much functionality I can’t help to play with all the time I doubt it’ll notice. It’ll more likely run down fast simply because I’ll use it more.

I’m also pleased that there’s no problem to just plug in the Magic to my Linux desktop and copy/sync the photos and the videos etc.

Google Integration

I realize some people will feel that the very tight integration with Google and Google’s services is a downside as it adds just another item that Google “owns” in your life. Still, it makes the experience very slick and as a user I get a lot of stuff “for free” as it just connects to lots of things that I already used and had accounts on. So gmail, sharing photos on picasaweb etc “just works”.

Decrypting ipods

Recently we’ve seen progress by the linux4nano guys in their quest to get custom code to run on an Ipod Nano 2nd generation. They’ve apparently managed to extract the bootrom off a 2nd gen ipod nano (my copy of their extracted data is here – a reminder on objdump usage: “arm-elf-objdump -D --target binary -marm [file]“). I believe their intent is to port Linux to the newer ipods. Possibly ipodlinux. They do mention providing the necessary info to Rockbox and yes we will welcome it.

A large crowd of Rockbox hackers have joined their IRC channel and have been hanging out with them and helped out discussing ideas and pushed them towards publishing their news and infos on how this all is accomplished etc. Their SVN repo hosts some (most?) of the tools made so far.

The Rockbox wiki page for nano2g has been updated and hopefully it will keep track of what happens.

There have been speculations, but I don’t yet know based on what facts, that this recent news and hacks will be usable on other recent (encrypted) ipod models.

Summary: very interesting progress has been made. Lots of it is still left to figure out. There seems to be a bunch of skilled people around and now we’re seeing information and documentation for this getting published so I can’t but to hope for a bright future!

Concepts of a new distributed build

It was time to make an overhaul of our distributed builds system for Rockbox. The one currently in place is quite fancy and it does build 106 builds in around 7-8 minutes, but during the years it has served us we have found a few areas where we want to improve.

The goals for the new system were primarily:

  • do all the builds faster
  • reverse the connection so that people can contribute clients easier
  • make a system that is more allowing for slower machines to contribute

The biggest weaknesses of the existing system:

  • The master uses ssh to the distributed clients, which forces them to have an accessible ssh server and port etc. It also makes it awkward for people behind NATs who wants to run more clients.
  • It only hands out a particular build to one client, so thus if a large build happens to get handed to a slow client towards the end of a build round, all the other clients will sit idle waiting for the last client to finish.
  • The build and the subsequent upload of results to the master are synchronous, so thus a client with a very slow uplink may spend a significant time on the upload before it can start the next build.

The  new system is currently in development. It consists of a server that runs on one of our main servers, and there’s a client script that each volunteer contributor runs on their systems.

The clients connect to the master on a dedicated TCP port, specifying user name, password, name of the particular client instance, what particular architectures the client can build and how many bogomips the client boasts. While bogomips is a bogus way to measure anything, we’ve started out using it for a rough way to sort the the build clients based on speed.

The clients keep connected to the server all the time. There’s a ping message from the master every N second of idleness to make sure the connection is kept alive. As soon as the master wants the client to do a build, it sends a message to it detailing exactly how it should build it and using what SVN revision. The client will then do the build at once, upload the results using HTTP to a dedicated place and then tell the server the build is complete.

The server knows about all builds to do at a  commit, what we call a build round. It has a rough “score” or “weight” for each build that grades them in a slow to fast order. When a build round starts, the server will first sort all builds based on number of times they’ve been handed out and as secondary sort key the “weight” of it. Then it loops over the currently connected build clients and hand out builds from the sorted build table. The server then continues to do that until all clients have three builds each to build. As soon as a build is reported to have been completed by a client, that client will get the next build from the sorted build list.

If a client connects to the server and the server deems the client to be too old (since it does specify its version in the handshake message), it will be told to update to a specific version instead and come back then. This way the server can update all build clients when important things are fixed.

The clients will soon start to get assigned builds that already have been assigned to another client. This is not a problem but in fact our intention. The client that completes the build first will simply tell the server, and the server will then tell all the other clients that build that same build that they should cancel that particular build.

A client that joins the server in the middle of a build round will simply get a bunch of builds immediately and join in. A client that disconnects during a build round simply won’t complete its builds and other clients will instead do them. The system is also tolerant against the fact that bogomips is lame to compare computers with, and that the build “score” may not be very accurate or even that some server will have very slow or very fast upload speeds at unpredictable times.

The build master itself does not know when to start a new build round. It simply knows about the concept and it knows how to tell clients to complete a round. To make the master to start a new round, you need to connect to the server’s listening port and issue a special command and provide a password and then you can tell the server to start a build of a specific SVN revision. Or to queue up a build to be performed after the current one if there happens to be one in progress already.

When a full build round is complete, a hundred or so builds have been done, and full packages and log files are now in a directory on the build server, the server will simply trigger an external script that then takes care of updating our build table etc. In fact, every single completed build will optionally trigger an external script to allow web pages or stats pages to get updated as we go.

This build system is currently pretty Rockbox-specific as this is the project and development system we’re writing this for, but there’s really nothing in this that must be this way. I’m sure that if someone (you?) wants to adapt this for another project, I’d be more than happy to assist and to help ensuring that this becomes a more generic distributed build system. Just raise your hand and step forward!

At the time of this writing, (primarily) me and Björn are still ironing out quirks in this new system to hopefully get it going live real soon…