1

Sorry for the long post. I tried to make it small but couldn't make it any smaller. Any help would be really appreciated. It would be great if you can explain from the basics.

In the book "Computer Networking A Top-Down Approach" the following is written: "One important characteristic of cable Internet access is that it is a shared broadcast medium. In particular, every packet sent by the head end travels down- stream on every link to every home and every packet sent by a home travels on the upstream channel to the head end. For this reason, if several users are simultaneously downloading a video file on the downstream channel, the actual rate at which each user receives its video file will be significantly lower than the aggregate cable downstream rate. On the other hand, if there are only a few active users and they are all Web surfing, then each of the users may actually receive Web pages at the full cable downstream rate, because the users will rarely request a Web page at exactly the same time."

I am getting totally confused by this, as I am having problems figuring out how exactly does broadcasting work in the case of cable networks. It is said that every packet sent by the head end travels downstream to every home. Now suppose the Cable Head End is transmitting a packet with, say, ContentX, and suppose 10 homes are connected to it. In this case 10 ContentX packets will be sent through the link, and each home will get one. But suppose 3 homes are simultaneously searching for Content1, Content2 and Content3 respectively, and the rest of the homes are inactive. In this case the Cable Head End will broadcast packets of Content1, Content2, and Content3 to each home. Which means, there will be 10 Content1 packets, 10 Content2 packets, and 10 Content3 packets in the network (as every packet is broadcasted to all users). So, shouldn't the downstream rate in this case be even lower (because there are more packets to be sent in the same available bandwidth, there will be congestion) than the time when everyone is trying to download the same video because, no matter what, the Cable Head End will transmit the video packets to every house? So, its better if everyone is asking for those broadcasted packets itself rather than asking for different packets.

I can feel that it is intuitively wrong to imagine that the downstream rate would be higher in case of many active users, rather than a few active users. But please help me figure out what is it that I am not getting here.

P.S.: In case of packet switches, if a packet arrives at a broadcast address, the switch makes copies of the packet and sends out through all the outgoing links. So, it there are many such broadcast requests arriving at the switch, the switch will have to make copies of packets for each of those requests and that would overload the network and ultimately reduce the content download rate at each host. This is what I think will happen. Why does the opposite happen in cable network?

2

Don't confuse layer-1 broadcast (sending a signal everywhere) to a layer-2 or layer-3 broadcast (sending frames or packets addressed for every host).

The original ethernet was on coax, a broadcast medium. Every host could hear everything on the network segment. That is why ethernet adopted MAC (Media Access Control). An ethernet host only listens to frames addressed to it (including unicast to its MAC address, broadcast, and multicast for groups to which it has subscribed), even though all the frames reach every host on the link. The medium has nothing to do with how the frames are addressed.

Think about your cable television. All the channels are being broadcast on the medium all the time, but the cable provider doesn't send each channel individually to all the homes; it sends each video frame on each channel once, and a television in any home can display it if it is listening to that channel.

The cable ISP also only sends a data frame once, the way it happens in ethernet. The shared medium may deliver it to all the sites on a segment the way cable video is sent once to all the homes on a segment. A cable modem only listens to frames destined to its specific address, ignoring all the others. The same thing happens when your cable modem sends something back to the ISP. The frame may received by all the cable modems on the segment, but only the ISP equipment is listening to frames addressed to it.

For your ethernet broadcast, a switch only sends the frame once to each switch interface. It doesn't make a copy for each host on the LAN, and it is no more onerous for a switch to copy broadcast frames to every interface than it is for the switch to switch a single frame to a single other interface. There can be multiple hosts on each switch interface (think hubs or switches connected to a switch interface). The problem with broadcast on ethernet is that it interrupts every host on the LAN, and it uses bandwidth on every switch interface. You may be thinking about a broadcast storm, where bridged (switches are bridges) ethernet has a bridging loop, so that broadcast frames never die, they just accumulate until the LAN collapses.

  • Going back to "The shared medium may deliver it to all the sites on a segment the way cable video is sent once to all the homes on a segment.". From this it seems like sniffing unencrypted traffic in your network segment would be trivial (assuming you used your own modem which listened to every frame). Nowadays I wouldn't worry about it, given the prevalance of TLS but I could imagine that at one time it was less prevalent. Would ISPs encrypt any of the frames it sends to all modems in the segment by default? Are there concerns about broadcasting this information? – Deepak Jan 10 at 17:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.