Future of DCTV

[converge]

Just reading through the FAQ and realizing that Defcon TV appears to fall by the wayside with the new digs at the Riviera.

Has anyone taken into consideration the possibility of live streaming or a chopped version of iptv to deliver similar content over the hotel network? Short range rf broadcast for the hotel? Is this simply its demise?

[Deviant Ollam]

i had heard that the tv network simply won't be carrying the talk feeds (the day of mythTV talk capping is dead, it would seem) but will be showing some hacker-related movies and filler material.

my dream setup for DCTV would be some sort of system where unused channels are opened up to whomever wants to broadcast anything... just supply either a box or a video system on a playback loop. imagine having hidden channels all over the places offering things like Futurama, Robot Chicken, and obscure anime 24/7.

there could conceivably be no limits on the content which one broadcasts, since a person would have to mod their TV's internal channel list in order to see things, thus eliminating the "but what if other hotel guests flipped around and saw farm animal porn" argument. if we had such a system i'd personally supply boxes continuously playing every episode of the above-mentioned shows as well as mythbusters, monty python, and the family guy. ah well, a guy can dream.

[TheCotMan]

Once we have the whole hotel to ourselves (not counting the casino) we may (guessing the future) be able to get access to more channels and if the convention is that big, maybe there will once again be room crowding issues that will force use hotel cable tv for presentations broadcasted to rooms.

Between hotel-room-tv talks and guaranteed access to actually be in a the room of a presentation, I'll take being in the room with the talk. (You can ask questions this way.)

[converge]

Doh, I just reread the FAQ and misread.. it does state that 2 channels will serve non-talk material. I still like the discussion in this thread though because I think a hacker conference could easily, unofficially pull off something like what Deviant mentioned, for the sake of coolness.

[Voltage Spike]

Has anyone taken into consideration the possibility of live streaming or a chopped version of iptv to deliver similar content over the hotel network?

How amenable would the parties involved be to this idea? I assume those running the cameras would allow us to use their feed again, but how would the network guys feel about us broadcasting a constant 7 Mbps of streamed data?

Are there any cheap or non-commercial solutions out there for broadcasting live video across a local network with error correction? I'm thinking along the lines of over-the-air digital television transmission (ATSC, ISDB) but using a subnet broadcast. I have seen very expensive products that perform this task, but it certainly isn't something you whip together on a weekend.

[dataworm]

TV? What's that? Oh yea my parents used to have one of those...

Why not pass to a new age and stream on internet?

[TheCotMan]

Network-based broadcast of video content sounds like a good idea, but there is one serious problem with this:
This would be DefCon: where all it takes is 0.02% (1 person out of 5000) of a attendees to ruin a network. (I'll be submitting a slogan like this to the DefCon 14 slogan contest btw.)

Though they work to keep the networks operational with clever application of skills, and use some clever technologies too, risk for loss of any service is always a problem.

If there is plenty of space in the rooms, then why not just go to the room of a presentation?

If the intention is to record the video, and it were broadcast on the network, would wireless do well with this, assuming even 1Mbps (combined audio/video) when you consider there may be 3 presentations going on at the same time, and people trying to download other stuff?

Use of broadcast address, or multicast capable/enabled networks may help to share throughput, but it may also increase demand of the network. Early manufacturer whitepapers papers suggested no more than 30 users per wireless access point.

If wired, then will you leave your wired workstation unattended when you go get food, go to the bathroom, or will you be anchored to the computer?

If rooms were part of the DefCon network (and had >= Gigabit, reliable, on-site-routing to the DefCon networks without leaving the hotel) then this might be an option. However, I don't recall any mention of any in-room network access being part of the DefCon network.

But hey! Maybe someone could volunteer. Maybe people will bring laptops with cameras and sit in multiple presentations and later broadcast what they capture, or build torrents. Then again, someone may encode HaXXXor Girls to be some bloated codec AVI, and share a torrent under the name "ReallyCoolDefConPresentationThatYouWantToSee. avi"

If you build it for DefCon, expect people to try to destroy it.

[bascule]

I think it's time for the ham radio nerds to bust out the SSTV. We can watch the talks in an awesome 12 seconds per frame

[Voltage Spike]

This would be DefCon: where all it takes is 0.02% (1 person out of 5000) of a attendees to ruin a network.

While true, it is also true that those with the technical know-how to shut down such a service would have little desire to do so.

(By the way, I like the translation of "it only takes one..." into a percent. )

Though they work to keep the networks operational with clever application of skills, and use some clever technologies too, risk for loss of any service is always a problem.

True, but we have already lost the service. Any uptime is an improvement over the current situation.

If there is plenty of space in the rooms, then why not just go to the room of a presentation?

As was mentioned in the previous thread on this subject, problems include conflicting timeslots, the desire to preserve the video, and the desire to just relax in a room after a rough night.

If the intention is to record the video, and it were broadcast on the network, would wireless do well with this, assuming even 1Mbps (combined audio/video) when you consider there may be 3 presentations going on at the same time, and people trying to download other stuff?

What is the data rate of the outside connection? If we assume 802.11b with a maximum throughput of a little over 7 Mbps (after overhead), then such a use would take half of the bandwidth. Of course, that is half to just one application...

Early manufacturer whitepapers papers suggested no more than 30 users per wireless access point.

Two early thoughts. The first is that such statements probably assume 2-way communication (31 transmitters on the same frequency) while IPTV involves only 1-way communication (1 transmitter). The second is that the users don't have to associate with the access point...

[TheCotMan]

While true, it is also true that those with the technical know-how to shut down such a service would have little desire to do so.

Heh. It does not take much knowledge to use up a lot of bandwidth, and there are "tools" that script kiddies can download to create DoS.
Skill to understand how to create a DoS tool is often inversely proportional to willingness to cause DoS.

Give a man a script and you'll have blue screens for days.
Teach a man how to script, and he is employed to stop blue screens with little time to create blue screens.

(By the way, I like the translation of "it only takes one..." into a percent. )

Thanks. It is right in-line with "how many nine's uptime" that I see in product ads.

True, but we have already lost the service. Any uptime is an improvement over the current situation.

This sheds light on an interesting thing I've noticed with people in hacking cultures: Reputation and Ego (not the same thing) tend to direct many people in the decisions they make. Some people with large egos may choose to not take an action because the risk for being publicly blamed might burst their ego-bubble. Some with reputations, may use their reputations as reasons to get hired for work. These people may be unwilling to even try things for fear failure in something new will cause their customers to irrationally assume failure of skill in one space means failure of skills in another space.
(Sort of an inverse to "If you know one thing, they think, you are skilled, two things, they think you are a master, but 3 things and they assume you know everything.")

Those not worried about failure are the ones most willing to try. For people without reputations, there is nothing but advantage to success, and little worry about failure.

As was mentioned in the previous thread on this subject, problems include conflicting timeslots, the desire to preserve the video, and the desire to just relax in a room after a rough night.

Well, there is always "The Sound of Knowledge" -- yeah, I know it is expensive, but it is an alternative.

Before there was video, sometimes my buddies and I would split our conferences and then meet and share.

I'm not against a return of video, but I am happier to hear about decreased risk for lines to see a presentation,a chance for no room evacuation between presentations, and plenty of space to not be rubbing sweaty elbows with people as freaky as me-- even if it means loss of video.

What is the data rate of the outside connection? If we assume 802.11b with a maximum throughput of a little over 7 Mbps (after overhead), then such a use would take half of the bandwidth. Of course, that is half to just one application...

Yeah. Um. With WiFi, I *think* (somsone can correct me) that the 1,2,4.5,11,22,54 Mbps advertised rates assume dedicated users. With more users, a time delimited multiplexing is used to share the throughput with many users, not counting an increasing overhead per user.

Two early thoughts. The first is that such statements probably assume 2-way communication (31 transmitters on the same frequency) while IPTV involves only 1-way communication (1 transmitter). The second is that the users don't have to associate with the access point...

With WiFi, there are differences from wired ethernet/10BaseT.
The Layer2 involvement is more complicated and I am sure I'll get something wrong here, but here goes:
With WiFi, association includes many things, one of which is an agreement on what "speed" is available for passing data. The closest similar technology might be the autobauding process that modems use to establish what "layer2" protocols to use, and what speeds to use. Both parties must speak to each other and agree upon speeds.
When streaming video, even if it were possible to use some sort of WiFi broadcast to all nodes, something would need to be done about selecting what packets to drop when users are far enough away, or in places with too much interference to keep associated at 11Mbps, but have to drop down to a slower data rate.
[Also, different speeds mean different signaling for bits... A unit set to receive at 11Mbps may not understand a 1Mbps signal, ulnless the 11Mbps unit was told to receive at 1Mbps.]

Wireless seems a closer match to simple hubs with shared access to throughput. A simple 10BaseT hub has (theoretically) 10 Mbps of throughput. Put 10 users on it, all trying to download from each other, and collisions will probably result, while each users ends up with less than 1Mbps for their download rate.

I can see use of broadcast to wired users, since the speeds are established with the layer conventions. Layer 2 switches can work well with giving each person "full speed" for a broadcast.

One of the wireless people here can probably give you more technical information on how throughput per user scales poorly between a wireless accesspoint and users as users increase-- even when trying to use IP broadcast/multicast.

[converge]

I'm not against a return of video, but I am happier to hear about decreased risk for lines to see a presentation,a chance for no room evacuation between presentations, and plenty of space to not be rubbing sweaty elbows with people as freaky as me-- even if it means loss of video.

Definately an added feature and getting in to see the show is definately a plus, considering that one reason for falling back on the video was because talks were simply too full, imo not a very good substitution although a nice consolation prize. That was not the only reason though. A couple times I have been literally stuck in my room, or Chris', debugging something for the contest and it was quite excellent to catch a little hacker jeopardy or dmca panel talk to clear my head of fucked up binary logs. Other folks have other reasons, as Voltage Spike pointed out.

Not saying its an end all must have, but a nice feature that I and many others have grown accustomed to over the past couple years ... one that seems feasible to continue considering the background many folks attending.

[Voltage Spike]

This sheds light on an interesting thing I've noticed with people in hacking cultures: Reputation and Ego

I'm not really sure where you were going with that, or how I helped you get there. In fact, I'm not sure I even agree, but it was interesting nonetheless.

Well, there is always "The Sound of Knowledge" -- yeah, I know it is expensive, but it is an alternative.

I wasn't going to be the one to bring up the DVDs ... or the media archives. Mainly because it isn't nearly as much fun to discuss.

I'm not against a return of video, but I am happier to hear about decreased risk for lines to see a presentation,a chance for no room evacuation between presentations, and plenty of space to not be rubbing sweaty elbows with people as freaky as me-- even if it means loss of video.

I'm obviously not going to complain about the new floor space. In any case, I've sat next to you before and it wasn't nearly as unpleasant as you imply.

With WiFi, there are differences from wired ethernet/10BaseT.

That is a good point I had not considered. It is entirely possible that access points transmit broadcast packets on a per-client basis, but I don't have the expertise to say. Some initial research indicates that broadcast data is indeed sent in a single frame without Layer 2 ACKs. As I understand it, the slower negotiated speeds are the result of breaking the data into smaller (Layer 2) frames (unlike modem behavior) so that even "slow" clients receive all broadcast data (albeit with a higher chance of corruption).

PS: I should add the standard disclaimer that I have never designed, implemented, or made an in-depth analysis of 802.11 hardware ... so don't use anything I say in your thesis paper.

[TheCotMan]

I'm not really sure where you were going with that, or how I helped you get there. In fact, I'm not sure I even agree, but it was interesting nonetheless.

It was meant to be another reason for people with knowledge and skill to not be motivated to offer such a service.

I'm obviously not going to complain about the new floor space. In any case, I've sat next to you before and it wasn't nearly as unpleasant as you imply.

Well that was Grifter's talk at DC 13, and there were plenty of seats-- we may have even have an empty chair or two between us, so you did not get the full experience of me and my sweaty elbows. (heh heh)

That is a good point I had not considered. It is entirely possible that access points transmit broadcast packets on a per-client basis, but I don't have the expertise to say. Some initial research indicates that broadcast data is indeed sent in a single frame without Layer 2 ACKs. As I understand it, the slower negotiated speeds are the result of breaking the data into smaller (Layer 2) frames (unlike modem behavior) so that even "slow" clients receive all broadcast data (albeit with a higher chance of corruption).

If this is true, and I have not checked on my own, that would seem to indicate that a broadcast in such a system would always be sent at the slowest rate of all associated access points/client sessions.
I just assumed longer periods for each bit-- a common practice for dealing with interference to communications, and as a method to run longer distances without interference damaging transmitted data with SCSI, Serial, Modems, and other datacomm devices.

[Voltage Spike]

If this is true, and I have not checked on my own, that would seem to indicate that a broadcast in such a system would always be sent at the slowest rate of all associated access points/client sessions.

I forgot that many of the IEEE 802 documents were free to download...

All frames with multicast and broadcast RA shall be transmitted at one of the rates included in the BSS basic rate set, regardless of their type or subtype.

I don't see the relevant section of the 802.11 document relating to the basic rate definitions, but I seem to remember them being 1 and 2 Mbps. This is certainly not promising, but remember that control frames, for better or worse, do not incur the switching overhead so the rates are "as advertised".

If I get some more time later I'll try to dig up any relevant research papers concerning video over WiFi.

[TheCotMan]

I don't see the relevant section of the 802.11 document relating to the basic rate definitions, but I seem to remember them being 1 and 2 Mbps. This is certainly not promising, but remember that control frames, for better or worse, do not incur the switching overhead so the rates are "as advertised".

This is a really bad sign, since transmitting with WiFi (be it broadcast or live sessions) is dedicated, and mututally exclusive to other transmissions.
If we assume 3 presentations with 1Mbps for each, then there is 3 Mbps and over some baseline.
If we assume the baseline is 1 Mbps, then that is one presentations and a DoS to all other users of the accesspoint for other purposes.
If we have 2 or more dedicated WiFi devices setup to broadcast the same presentation, there is risk for interference between these, and when adding more presentations-dedicated WiFi, even more interferences, and adding WiFi for web browsers/games, perhaps more interference.

If I get some more time later I'll try to dig up any relevant research papers concerning video over WiFi.

Cool deal. It will be interesting to see what is found.