Become a fan of Slashdot on Facebook

 


Forgot your password?
Close
typodupeerror
×
Music Media

AOL Developing Cheap Switch for Audio Streaming 166

Posted by timothy from the when-banks-compete-you-win dept.
legaleagll writes: "According to a Fortune magazine article and a follow-up article on ZDNet, AOL is developing a cheap switch that can handle streaming audio for 10,000 users, versus current technology of 100 - 1,000 users per box depending on expense of system. The code name for the product is Ultravox and was apparantly spurred into existence because RealNetworks is now offering internet service for cheaper than AOL. I'm a little skeptical because I'm not sure how the use of an intelligent router would eliminate the need for the expensive systems to stream the audio. Wouldn't moving the software for streaming onto the router make for a more expensive router and still require the expense a box outside of the router anyway?"
This discussion has been archived. No new comments can be posted.

AOL Developing Cheap Switch for Audio Streaming More

AOL Developing Cheap Switch for Audio Streaming

Comments Filter:

  • The 80's live on (Score:4, Funny)

    by cryptogryphon ( 547264 ) on Thursday June 27, 2002 @03:07PM (#3781106)
    I think Midge Ure might have something to say about that!
    • My thought exactly...

      "Dancing with tears in my eyes..."

    • This means nothing to me!
  • Yeah but given the fact that serving that many users is now *that much* more expensive, this seems irrelevant? Not to mention how much dedicated bandwidth *still* costs nowadays (no, not cable, but the stuff you use to host 10,000 users....)

    -kwishot

  • Not sure how this will help. (Score:5, Insightful)

    by gorilla ( 36491 ) on Thursday June 27, 2002 @03:07PM (#3781108)
    The problems with streaming don't seem to me to be the capacity of the streaming boxes, but the bandwidth and legal problems. Faster hardware is nice, but the streaming costs are still going to be too high.

  • I wonder... (Score:1, Redundant)

    by Johnso ( 520335 )
    I wonder if they'll prepend every song coming through with "You've Got Music".
  • you only stream one band! [allmusic.com]

  • The ultimate conspiracy.... (Score:3, Funny)

    by Dark Paladin ( 116525 ) <jhummel.johnhummel@net> on Thursday June 27, 2002 @03:10PM (#3781145) Homepage
    Phase 1: AOL develops the Super Audio Switcher 2000 - let your whole business listen to the music you would normally pipe through the company speakers at once! Want cool hold music? Just plug the SAS2000 into your PBX!

    Phase 2: Hello, we're lawyers from AOL. Seems that your business has been streaming Time Warner music without our official permission. You now owe use $1,000,000,000 dollars - or your entire company, whichever is greater.

    (Please note: the above is "sarcasm", and is not meant to reflect a real world situation. Please turn "cluelessesness" switch to "off" when reading this post. That you.

    • Re:The ultimate conspiracy.... (Score:1, Informative)

      by Anonymous Coward
      Please note: the above is "sarcasm" ...

      Actually the above is satire. I would not qualify it as sarcasm since it's lame. Hell it's not even funny satire (despite what the mods think).
    • Phase 3: Profit

      I have to admit, you've finally come up with a scheme wherein phase 2 is known up front... congratulations!

  • Multicast? (Score:1, Interesting)

    by Anonymous Coward
    It seems to me that moving the streaming onto the switch refers to multicast...At that point the bandwidth issues for supporting 10,000 users become much more manageable. Now, just have to find those 10,000 users on multicast enabled networks...

  • Effectiveness (Score:5, Informative)

    by kwishot ( 453761 ) on Thursday June 27, 2002 @03:13PM (#3781173)
    "I'm not sure how the use of an intelligent router would eliminate the need for the expensive systems to stream the audio. Wouldn't moving the software for streaming onto the router make for a more expensive router and still require the expense a box outside of the router anyway?"

    I'm assuming that for these big operations that have multiple sites for streaming, it works something like this: You have a source for the stream, which no end-users ever touch. The job for this machine is to feed the routers, which can be in various areas of the world or whatever. That way each router has a dedicated "in" stream so that it can feed the masses. In a smaller operation where multiple sites aren't needed, this piece of hardware isn't needed either, so this thing actually does save money for an establishment that can take advantage of it's potential.

  • Why they want to do this (Score:5, Informative)

    by Smallpond ( 221300 ) on Thursday June 27, 2002 @03:13PM (#3781177) Homepage Journal
    "I'm not sure how the use of an intelligent router would eliminate the need for the expensive systems to stream the audio."

    Currently 1000 users == 1000 streams. An intelligent switch looks at the content, says broadcast one stream to the 1000 users. Sounds simple. Not, though. Check articles on "Layer 4" switches.
    • You should also check out the concept of IP Multicast. Supported by modern routers, it could significantly help streaming type of applications.

    • Re:Why they want to do this (Score:4, Funny)

      by Washizu ( 220337 ) <bengarvey@co m c a s t . net> on Thursday June 27, 2002 @03:26PM (#3781287) Homepage
      "Currently 1000 users == 1000 streams. An intelligent switch looks at the content, says broadcast one stream to the 1000 users. Sounds simple. Not, though. Check articles on "Layer 4" switches." I thought you weren't supposed to cross the streams?
    • Currently 1000 users == 1000 streams. An intelligent switch looks at the content, says broadcast one stream to the 1000 users. Sounds simple. Not, though. Check articles on "Layer 4" switches.

      congratulations, you just described multicast.

      Multicast has been support in routers and OS's for years. No new propreitory systems needed.

    • "Layer 4" switches

      ps there's no such thing as a layer 4 switch. This might be some new market-speak, maybe.

      Network switches operate at layer 2, eg. ethernet, etc. If you at layer 3, then you're routing.

      layer 4 to 7 filter is new, but they're definately not switches. Most times AFAIK, they're called 'packet shapers'

      • For an article on layer 4 switching, see
        Layer 4 Switching [networkcomputing.com]

        Layer 2 - LAN layer - switch based on MAC address
        Layer 3 - Network layer - switch based on IP address
        Layer 4 - Transport layer - switch based on contents (like TCP port number)

        Although technically they aren't switches, they're called switches because they do layer 4 switching, see?
        • For many vendors, Layer 4 switching simply refers to applying filters to IP packets based on Layer 4 information. For example, if you want to filter all Web traffic off certain ports, or all PointCast traffic from your network, a simple Layer 4 access-control filter will do it. In a way, these products function as rudimentary firewalls, except they aren't designed to alert you when a user violates a rule--they simply enforce it.

          Thanks for the link, it actually cleared up what I was trying to say.

          I consider that class of device a "filter" or "traffic shaper" rather than a "layer 4 switch".

          But yeah, you point is valid. I may as well surrender the fight and call them "layer 4 switches" like everyone else :) sigh.

      • Network switches operate at layer 2
        Actually, "switch" is market-speak. They are actually multi-port bridges that may or may not be bridging between unlike topologies (ethernet to token-ring, etc.). But, see, bridges were typically 2-ported. And they were often used to "bridge" between unlike topologies. When someone came up with the idea of making a bridge with more than 2 ports that would bridge between multiple wire segments even of similar topologies, well...it was just easier for marketing to call it a "switch" instead of give a big, long description of what it actually is...a really cool, tricked out bridge.
        • it was just easier for marketing to call it a "switch" instead of give a big, long description of what it actually is...a really cool, tricked out bridge.

          true.

          Nice point, btw.

          But get this. I work with foundry bigiron 8ks. These can inpect layer 4+. Are they "layer 4 routers"? :)

          If I have my dns server send certain queries to other dns servers, is it a "layer 7 router"? I mean, it is routing at the 7th OSI layer.

          if we look at this conceptually, switching and routing much alike. But we call them different things, making it easier, in part, to communicate what layer we're 'thinking' in. ie. encapsulation.

          to be honest, I'll admit mine is not a very important or particularly insight argument. but hey, it's slashdot and I'm bored :)

      • You better check again. There is a wonderful world of switches out there now. Every kind of switch from layer 2 through layer 7 [nortelnetworks.com]. There are many other manufacturers offering similar products too.

        The difference between these "switches" and the layer 3 routers that you are use to is the hardware architecture. While older routers consisted of routing code on a CPU passing packets through the CPU and out the appropriate interface, these layer 3-7 switches use ASICs(Application Specific Integrated Circuits) to process the packets rather than the CPU. This translates into a MUCH higher throughput than the CPU based router could ever handle.

        If you get in medium to large networks you will see these switches everywhere. Switches with 128 or more gigabit ethernet ports all switching at line speed. Your Alcatel DSL router or Cisco 2600 doesn't even know what line speed is.

        Indeed, there are layer 7 switches. Some of the Nortel/Alteon switches can even switch based on URLs. Makes for a much faster hosting site.

    • well thats not a stream, thats a brodcast. you would have to assume that all those users asked to connect to the file at the *same* time, which im sure does happen a lot, but cant happen all the time.

      the best way for it to work would have the router cache the file in its own ram, or at least a part of the file to help internal network traffic...but once it got to the net it would still be a huge bandwith hog.

  • How this might help. (Score:3, Informative)

    by Christopher Thomas ( 11717 ) on Thursday June 27, 2002 @03:15PM (#3781194)
    I'm not sure how the use of an intelligent router would eliminate the need for the expensive systems to stream the audio. Wouldn't moving the software for streaming onto the router make for a more expensive router and still require the expense a box outside of the router anyway?

    My guess? This is a smart cacheing system.

    The ZDnet article makes much reference to bandwidth congestion as a major stumbling block to streaming media. As a user will always get the same data back when requesting a media stream, you could set up smart router/proxies to cache, oh, the first few blocks of all media clips and the entire contents of the last N requested clips (up to a space limit). Assuming many users are listening to the same clip in the same general timeframe, you only have to stream the clip to the cacheing router once (instead of many times), saving bandwidth and load on your source box. Non-media traffic is routed as usual.

    Or this could be something completely different. But a smart router/transparent proxy type of deal seems to be the most beneficial thing to develop.
    • If you do this, then maybe the RIAA wouldn't be able to charge you the full amount for your broadcast. If all streams that end up in AOL subscribers go through a handful of these routers originally, it might look like only a handful of listeners are actually tuned in to your station.

      Or we could just challenge the legality of the rates, that'd probably hold up better in court.

  • Multicast anyone? (Score:3, Insightful)

    by JohanV ( 536228 ) on Thursday June 27, 2002 @03:16PM (#3781205) Homepage
    Why develop propietary switches and fileformats when all that is required is a full implementation of multicast (which is just a part of IP)?

  • MBone (Score:4, Interesting)

    by Mr. McGibby ( 41471 ) on Thursday June 27, 2002 @03:17PM (#3781210) Homepage Journal
    Everytime I hear about streaming media again, I think back to the mbone. Why give a stream to every single user when you can intelligently stream media using the very thing that makes the internet what it is, it's ability to route packets to their destination? Why should I have to send out 1000 copies of the *same damn thing* over my wire when I could just send one copy and let the routers send copies to subnets that are going to use it?

    Whatever happened to the mbone!?!?!?!?
    • Yes I seem to remember watching live video on the mbone around 1994. And then *poof* no more mbone. At one time there was a host mbone.yahoo.com, but no more. The old mbone information web at www.mbone.com is squatted.

      It's too bad really. Multicast is neat.

      • The MBone is alive and kicking. I am having 79 channels in my IP/TV listing as we speak (while watching an MPEG-1 real-time transmition of a TV channel).

        Best site to get more info and downloads is http://videolabs.uoregon.edu/
      • Normally, I would suggest that you try here:

        http://web.archive.org/archive_request_ng?collec ti on=web&url=http%3A%2F%2Fwww.mbone.com&Submit=Take+ Me+Back%21

        Unfortunately, The Wayback Machine seems to be having issues.

        Maybe Google has a cached page of this cached page.

        ;^)
    • Something else to check out:

      OpenMash [openmash.org]

      This is where I used to work at Berkeley. The Berkeley Multimedia Research Center [slashdot.org] was big on the MBone for a while.

    • MBone only worked so well because organizations were willing to sacrifice tons of their bandwidth so that the sender didn't have to use up much bandwidth.

      What you most likely want is Multicast on the internet. The problem with that? Security. Right now people take over hundreds of boxes with big pipes and flood Yahoo or Ebay. Just imagine if each one of those boxes could send out each one of those packets to 16 million hosts! A MDDoS could be the very thing to actually bring down the entire internet.

      So, the idea is cool enough, but security needs a great deal of consideration. So far, I haven't seen any Multicast router implimentations that do all the advanced things it would need to do, to prevent hosts from getting data they don't explicitly request.

      I would also love to see Multicast on the internet (multicast Slackware ISOs... Sweet.), but nobody has stepped up to develop it as a viable option for untrusted users on public links.
  • There are about 50 companies that already make this product. For a little service called content delivery networking.

    cisco, cacheflow, network appliance, volera, and the list goes on and on
  • competition is spurring innovation, and creating something that will be a benifit to the end us^H^H^H er.... wait a minute... this won't benifit those who streaming, just who server up the stream... d'oh!

    "Wouldn't moving the software for streaming onto the router make for a more expensive router and still require the expense a box outside of the router anyway?"
    • But by helping those that serve, they in turn give us more (different) feeds and more (different) content. Well I hope :-D Ohh and maybe they won't declare bankruptcy because they use 100x more bandwidth than they can afford.

      -- DeionXxX

  • Multicast/unicast conversions (Score:4, Informative)

    by silas_moeckel ( 234313 ) <silas.dsminc-corp@com> on Thursday June 27, 2002 @03:21PM (#3781244) Homepage
    While niether article has anything technical I have worked in this industry for awhile now. I'm going to assume some things first off this is primarly for live streams only or simulated live as a router has no place mucking with vod streams it dosent have the memory to help besides some QOS things. OK now assuming it's live the best way to get the data out is multicast it but unfortunatly the Tier 1 ISP's cant figure out how to bill for Mcast and it dosent reach far enough like past an ISP and nearly never to anything like dialup or broadband it's a college toy. Now insert a router that can take multicast live streams and do a unicast conversion (it's not much harder than nat and just about everything that can load Cisco IOS can run NAT) Now that changes the math on the head end the servers only need to support the stream 1 time and the additional connections for authentication (non scalable multicast as real would term it) this would be a wonderfull thing to the source people and the end user (especialy if the cable head ends can convert back to multicast) and AOL has the size to get the teir 1's to play ball AOL is a huge installed base.

    But like I said this is all conjecture on my part. It may just be them whalking the arrowpoint boxes again they are routers of sorts (load ballancers realy) that can accept drive space.
  • I daresay that I am in strong favour of any technology that can increase the audiences of struggling musicians. People forget that the current wave of civilisation began in the Rennaisance, and was due largely to the activities of a few visionary artists, and musicians in particular.

    The bulk of our energies should go into supporting culture and the Arts. Otherwise, we end up no better off than our ancestors in the Dark Ages. And that, fellow Slashdot readers, would be a tragedy indeed.
    • " Otherwise, we end up no better off than our ancestors in the Dark Ages. And that, fellow Slashdot readers, would be a tragedy indeed."

      I don't think the Slashdotters need to worry, none of them use AOL. :)
  • For live events, IP multicasting seems to be the way to go, but where is it?! I was using the MBONE back in '94. ISP's have never configured their systems for IP multicasting. It seems to me that IP multicasting would save a ton of bandwidth!

    BTW, my understanding is that Akamai internally uses IP multicasting to send data out to all it's caching boxes. Now why couldn't they get ISP's to multicast enable their networks? It would save everyone some bandwidth.
    • Yes it would save them bandwith ISP's make money on bandwith thus bandwith is scarce as it cost money. Multicast is a great thing and I beleive it's REQUIRED to work in IPv6 and thats one of the reasons it's taking so long. Nobody wants a user with a t1 connection for 1k a month to use up say 30mb a sec in peering bandwith that could cost that ISP real money if they are a tier 2 or below. We as end users want Mcast the streamign peopel want Mcast the top ISP's dont want it going anywhere past there boarders (you cang et mcast connections ot top ISp's it just wont go anywhere from there)
  • This is just another futile attempt from AOL to grab on to the and I quote "millions and millions" of people who got wise to their crap.

    • This is just another futile attempt from AOL to grab on to the and I quote "millions and millions" of people who got wise to their crap.

      That's "billions and billions" and if you're going to quote Carl Sagan like that, you should at least give him credit!

      oh, wait a minute... :)

      GMD

  • Remember that, depending on the format, audio streaming is more complicated than what can be handled by mere multicast. The server must splice in headers containing codebooks for each new connection.

    What you could have is an intelligent server/router system wherein the router handles accepting new connections then after the codebooks have been sent adds the connection to a quasi-multicast group. In this way, the server would only have to handle new connections and continuing to feed the router with the source stream.


  • . . .it would be built out of a hacked XBox.

  • It makes sense AOL has shoutcast and winamp (Score:3, Insightful)

    by pcx ( 72024 ) on Thursday June 27, 2002 @03:29PM (#3781321)
    AOL already has the broadcasting infrastructure and the reciever infrastructure. Shoutcast [shoutcast.com] uses streaming mp3s which do stream very well considering its not the best compression out there anymore. Winamp [winamp.com] is probably still one of the best mp3 players out there.

    The problem of course, if they want to kill real audio, is that AOL does NOT have the infrastructure to do video. Others might say the problem is that all this is bottled up in the mess which is currently time-warner. The company which bought up all the tools [netscape.com] to destroy microsoft then failed to develop them.

    But all this is a moot point because a sizable percentage of people haven't used any real audio product or service in ages because microsoft's media player is adaquate, free, and doesn't require installtion (since it's bundled with the os) or navigating past the "pay for something you already have!" screens to get to the "free" player.

    Real audio is a dying format. All AOL has to do is either buy the company (doubtful given the current economic climate at time warner) or simply throw their support to quicktime or windows media. Eventually real audio will go the way of so many other dot.coms that tried to play in microsoft's sandbox.

    • I'd have to agree with you as far as Real's future, except for one thing:

      Real's compression schemes for streaming media currently top anything that Quicktime or Windows Media have developed... specifically in the area of voice recording. Their ability to preserve the nuances of dialect and inflection top anything else out there, all while creating a file nearly 1/3 smaller than comparable Quicktime compression.

      Just a little FYI for anyone who uses compressed audio for streaming applications.
  • I can see why AOL would want to do this, seeing as they seem they provide the bandwidth (seemingly for free) to some big streamers like digitallyimported.com who have thousands of unicast streams at 128Kbit. Never quite seen whats in for them myself, but kudos to them for doing it anyway.
  • Can you imagine? What next? Cars? "You've got gas!"

    Do we honestly need ANOTHER revenue (pardon the pun) stream for AOL/Time Warner? IF this is going to keep happening, why don't we just cut to the chase and impose an AOL/Time Warner tax on everything we do? Then they can just sit down and shut up...

    In the mean time, I really think someone ought to take AOL's attempt at cornering the VOIP sector and squash it like a bug by producing one Killer Router/Switch. Sounds like a job for Cisco.. (not to be confused with Sisquo)
  • ah... Vienna

    (Ultravox!)

    (The new wave band).

  • ..will they be dancing with tears in their eyes?

  • You've got sound! (Score:4, Funny)

    by MH ( 25322 ) on Thursday June 27, 2002 @03:46PM (#3781461)
    Feel sorry for the network guys...having to listen to "You've got sound!" all day from one or more of these routers would probably get rather annoying...

  • Reality is: (Score:2, Funny)

    by Anonymous Coward
    AOL users are morons anyway and would be WAAAYY too stupid to know how to use this technology.

    Am I right on this? HrmmmmMM?
  • Today America Online corperation announced a new process for streaming full motion video and full 2-channel stero sound to 10,000 simultaneous viewers. The technology was discovered in the back room of an Time Warner office during the merger last year.

    Its called cable fucking T.V. what the hell is this crap?

  • The expense is Real Audio and bandwidth (Score:4, Interesting)

    by AIXadmin ( 10544 ) on Thursday June 27, 2002 @04:03PM (#3781565) Homepage
    Last I checked the two biggest problems with massive streaming audio were 1) the high cost of a real audio license, and 2) the high cost of bandwidth. The latter might not be a problem for AOL . But the former is.

    AOL could be writing this switch to use MPEG-4 which would solve both problems to some degree. Consider also how much money AOL has to put into this project. The ROI could be huge.
    • I worked on the west coast with the Netscape Learning webcast group. We handled at one time, about 15,000 users. there's realserver drops all over the country on AOL's backbones. It was really fun watching the machines that were hosting our irc server and web servers (for Q&A) hit a load of 100% in about 20 minutes :P

  • IANASE (Score:3, Informative)

    by SkyLeach ( 188871 ) on Thursday June 27, 2002 @04:25PM (#3781719) Homepage
    I am not a streaming media expert but I do know routers and gateways.

    This could speed up streaming media because the bigest limitation on SM is handling very large IP stacks of high-bandwidth connections. Dedicated hardware which does not have the overhead of OS and general-purpose software can handle a lot more connections. In addition, it would probably run cooler and fit into a 1U rackspace.

    Those features should make it desireable to any ISP who provides streaming media.

    Throw in 4GB of ram and each 1U server could probably handle most if not all of the average ISP's SM content cached.

  • Number of users (Score:3, Interesting)

    by pjrc ( 134994 ) <paul@pjrc.com> on Thursday June 27, 2002 @04:30PM (#3781762) Homepage Journal
    can handle streaming audio for 10,000 users, versus current technology of 100 - 1,000 users per box depending on expense of system

    Apple claims that their QuickTime Streaming Server [apple.com] can send 4000 silultaneous streams. That's a lot more than 100 to 1000.

    It's also available as an open-souce project [apple.com], depending on your exact definition of open-source (not Free for all uses, apparantly).

    Wouldn't moving the software for streaming onto the router make for a more expensive router and still require the expense a box outside of the router anyway?"

    Yep, but if the box is streaming the same exact stream to lots of users, it could save a whole lot of expensive bandwidth by transmitting one copy of the stream over the long-haul backbone lines, where presumably a switch nearer to a cluster of users could transmit individual copies to the users over the "last mile".

    I don't know if that's what they're really doing, but it'd be the smart thing to do. Bandwidth on the internet backbone is a lot more expensive than servers and switches.

  • great timing by aol (Score:2, Insightful)

    by binarybum ( 468664 )
    guess this development couldn't have come at a better time huh aol? Now that streaming radio is becoming reduced to clearchannel style crap packed full of advertisement and delivered to you by Microsoft, Disney, and yes now TW/AOL, it would seem appropriate for them to devolop propritary technology that prepares them for this vast new market they've conqured by slipping greenbacks into the RIAA's panty straps.
    btw: they're out to make sure you're not exposed to new music at work too: http://www.msnbc.com/news/773100.asp?0si=-

    listen to clearchannel.... by cd's for $15 a piece from the top 40 rack at Sam Goody... now dance... dance... dance...
  • See subj for translation from journalistic into english...
  • There are situations where increasing the bandwidth in a network can actually decrease performance. say you have some routers working hard to keep up with 10Mbit links.. you say.. "sure wish my net was faster" so you upgrade to 100Mbit. Now the routers are freakin out and dropping everything left and right and throughput is actually going down instead of up.. and now your fired too.

    Better hardware and better protocols (ATM comes to mind) can go a LONG way in stopping jitter (variance in delay) and decreasing delay. I think this is AOL's thinking with this new switch.
  • One of the big problems with most streaming audio of which I am aware is that it uses a TCP/IP connection from the listener to the sender. If I listen to Radio Paradise [radioparadise.com] and my wife listens to it in the next room, the bits come down the wire all the way from Radio Paradise *twice*. This means that unlike with radio, webcasters' bandwidth usage is proportional to the number of listeners.


    With multicast, the distribution process is spread out across routers, so it's much less likely that the same bits will cross the same wire twice. If multicast routers were ubiquitous, the same bits would never cross the same wire twice.

    If you think about it, this is a fantastic deal - one of the big problems with broadband right now is that the broadband provider has a nice fat pipe to each subscriber, and a similarly fat pipe to the whole rest of the world. So there's a lot of contention on the pipe to the outside world, whereas the pipe to your house is mostly idle. With multicast, a 192kbps feed down the pipe to the outside world can put 192kbps on a significant percentage of the customer pipes. So the ISP can feed much more data to the subscriber for much lower cost.


    Interestingly, this also works to the webcaster's advantage - if you want to set up a webcast service on your home machine, it's no skin off the ISP's nose, because you're not pushing 192kbps *per listener* out their pipe to the world - you're pushing a total of 192kbps, which is much less expensive for them. This reduces their incentive/ability to provide lopsided connectivity, where you can receive a lot but send only a little.


    My only worry is that AOL might be deploying something proprietary and non-interoperable, and then the ISPs will wind up paying for a system that we can't use this way, even though it would have cost the same as a system we could use this way.

  • For true live or "like live" (c'mon you _know_ what I mean), surely the best way to deal with this stuff is to send one packet that will eventually reach every node that has declared interest, ie true multicasting?
  • We still need some techical information to determine if it's a good device.
    When talking about streaming audio/video, you can do 2 things.
    1) If it's a recorded stream(not live), you can cache the stream it to the users, like a HTTP proxy
    2) If it's a live stream you don't cache it. You split the live stream, which means the the box recieves one stream(or two to prevent blackout/ensure quality) and then split it to as many as the box can handle.

    There are a few of these types boxes available today, some are kinda reverse engineering the protocols. Those that do it properly, reports back the number of streams that the serve to the origin server(fx. Real). The broadcasting server then sees 2000(or whatever)clients attached but only the bandwith of one(pr. streaming speed). This means that copyright and licence can still be monitored by the broadcaster.

    There are many uses for this. One is companies that allow people to listen to net radios etc, watch streams to save outgoing bandwith on their internet connection. Works well here, only problem is with broadcasters that don't allow proxies to split/cache their stream because of some of these boxes, as I mentioned, have rev.eng. it so they don't care about reporting back the streams they are serving and the broadcaster needs to know how many because of copyright issues.
    Another usage is that when you got your 100's of servers running. It requires much less maintainance to have these boxes running than adding another server to serve streams.
    Big worldwide companies that does broadcasts internally(education, pep talks) has a box on each location to save traffic on their WAN.
    Then there is the ISP's that can have these boxes on their POPs. In the age of xDSL, they can put a box on every location where they have their xDSL housed. This helps, again, on the bandwidth on their backbone network. The question here of course is if there is money enough to be saved/earned from this. Those that would really benefit from this would be the broadcasters. Will the ISPs install these boxes so the broadcasters(customers) can save money? Maybe they would need to sign a contract before their steams would be cached/split. And then there's always Akamai. We have seen with the reality shows on TV, many users running streams from the show. Even though these shows have peaked, I'd say that there's still a lot of potential in streaming services.

    Ok I could go on for hours about this. :)
    What I see as the biggest challenge is to have all the broadcasters to allow proxies to cache or split their streams. All to often they just deny it because they are concerned about loosing control about the number of viewers, specially on pay per view services. I think we don't really need another cheap box, we need to ensure that these boxes will be used and in order to do this, the trust between the ones making the boxes, those who install them and the broadcasters, are established. Otherwise the boxes will have only be used by the broadcasters themself and if they make deals with ISP's to split/cache their streams, which is a shame. Imagine a internet where every POP has a streaming box installed. The bandwidth recieved by the users in the home could be huge with only little load on the backbone / internet.
  • I know this is off topic, and nitpicking, but this has bugged me for awhile and now is my chance(no offence meant to the story poster).Whenever I see a post about a certain thing and it's a very detailed description of it and it is called by a 'codename' isn't that...well, stupid? If it is referred to by a codename then we wouldn't know what it was, that's the reason for the codename to hide it or to confuse a competitor. Isn't it just a name and not a codename?

  • Peer to peer streaming (Score:3, Informative)

    by Sadiq ( 103621 ) on Thursday June 27, 2002 @05:57PM (#3782339) Homepage
    Some of you guys might want to look at VirginRadio's website (www.VirginRadio.com), they have a broadband stream (you've got to put in a password to access it, free reg though). The twist is, it's a peer to peer stream, they use a plugin by AllCast.

    Apparently, what this plugin does is contact the server and obtain a list of listeners, it then find a listener with atleast a 100kbs upload capacity and connects to them, receiving their stream from them. You'd think there'd be lots of problems.. e.g, it'd be slow, you'd get your bandwidth sucked away, you'd get cut off randomly. Well.. no. I've been listenening for about three days now, i've been cut off only twice in that time and i've had somebody else connected to me most of that time as well and I didn't even notice it till I looked at my upload stats later in the day.

    I'll put my money on this being the way to get around the problem, why does the server need to be the one to be dishing out all the bandwith? Why doesn't the server just serve the streams to other users, who in turn serve it to other users. Fairs fair.
  • If there was a way for the AOL system to act as a peer-to-peer system, then once a user (A) has begun caching the stream, another user (B) could start streaming the cache from user (A). That way, the router itself would not have to do too many simultaneously, and the great majority of AOL users would actually be streaming from someone else's cache.

    I know, lots of potential pitfalls.

    But since AOL already installs its software on user's computers, and a lot now are starting to be on cable modems, this could work.
  • check out: http://www.surgientnetworks.com
  • Similar concept, but already exists, and does the same thing with video.
  • those of us that have worked for DSL/Cable providers and have had the tech support calls about AOL's Software/client (bloated mess of code) screwing up the TCP/IP and network settings, you would also know how intrusive and controling it is on a typical PC.

    Earlier someone said something about running NAT or a form of NAT on the new SM "router".

    If you think about the fact that to use AOL's service you MUST have their client. They could very easliy have 1 or a few of the SM routers at each POP as needed. Set up one or more servers to deliver SM content only to the SM routers and run NAT to each (AOL Connected computer) behind it, thus reducing the effective WAN traffic from 5000 streams to one or two for each POP.

    How will the router integrate with the desktop PC?

    Think about a VPN with out the encryption, where the router is the remote network and the AOL client (PC) is the host. With out the need for encryption the processor overhead is limited.

    Comming soon to a DELL near you!

    "AOL 8.1 totally redesigned for streaming media."

    AOL acomplishes this-

    1) New users brought in by the consept that they too can get [decent] streaming media off their 56K.

    2) Old users remain hooked- it actually works, and a ID10T can opperate it.

    3) Finally while we would think bandwidth is not that much of a issue, it would still cut in to their margins, and thus the say 5+ million spent on the project development would prolly get made up in the reduced infastructure demands, reduced bandwidth utlized (thus no additional hardware for some time, outside of this project of course) and less human/shell script overhead as far as maintaining a decent QOS.

    My $0.04
  • AOL's so easy to use, no wonder it's #1!

Slashdot Top Deals

Software production is assumed to be a line function, but it is run like a staff function. -- Paul Licker

Close