Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : AUTOHOME, Inc -- Ignore unavailable to you. Want to Upgrade?

To: E. Davies who wrote (4882)	1/28/1999 11:36:00 PM
From: ftth	Read Replies (3) \| Respond to of 29970

re:<<Does anyone know if cable modems are running real-time compression algorithms?>> What, 30-40 Mbps in a 6 MHz bandwidth isn't enough for 'ya? :o). Actually, for video, MPEG2 compression will be used initially. Most deployed modems are data-only though (for now, anyway). For data, no compression is used (in the sense that I think you mean anyway), but the need hasn't really been there yet. It doesn't always gain something though (e.g. download a zip file--doesn't compress much more if at all). If the need came up, sure, it could be applied. But for multiple small transfers it may not in general gain much net-net because the packet may just get padded out to fill a frame (which it may have done anyway, without compression). As far as the multiple access scheme, remember this only applies to upstream transmission in the 5-42MHz range, or there bouts. The MA scheme depends on the implementation, and most deployed systems are proprietary. For standards-based deployments (i.e. DOCSIS), the upstream is a mixture of contention and reservation-based transmission "mini-slots." The function of the slots is dynamically controlled by the headend, and each modem is synchronized to the headend, so it knows which slots are designated for what purpose and/or designated for specific modems, and can act accordingly (or request slots to be allocated for a specific purpose, i.e. reservation). It also supports variable length packets to improve efficiency. The ranging process that each modem goes thru at initialization (and periodically thereafter) allows the delay times to the headend to be known by the modem. From that the modem knows how to offset the sync signal (the branch to each modem is a different delay to the headend so it has to offset it's view of synchronization to align with the rest of the node). More than you probably wanted to know. dh

To: E. Davies who wrote (4882)	1/29/1999 12:19:00 AM
From: George T. Santamaria	Respond to of 29970

<I got the 4Gb number very crudely. I assumed you could fit 100 6Mhz channels on a coax line each one providing 40Mb. Im sure that Dave Homes calculation of 2.6Gb is based on a lot more detail than that....> The real limit per channel is set by the s/n ratio established by network standards to be around 56Mb/6MHz~ 9.3bit/Hz, one-way. Modern HFC networks are about 785 MHz total BW and next generation networks will be at 1GHz. Therefore, the current generation network will have ~7.3Gbit ultimate BW and the next will have ~9.3Gbit. That ought to be enough MPEG content for ~250 homes w/ 2 TVs each, occupied with 2 couch potatoes per home, at once! The more difficult aspect of system capacity is providing 2-way services, such as cable telephony. This, along with internet access will crowd the lower 50 MHz of BW which is assigned for the uplink. SFA, for example, reccommends the network planners put no more than 250 homes to a node, primarily because of this consideration. These huge bw's will be realized when analog TV channels are dispensed with and newer generation modems, such as those using BRCM chips become commonplace.