SI
SI
discoversearch

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 : Voice-on-the-net (VON), VoIP, Internet (IP) Telephony

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: Stephen B. Temple who wrote (1648)10/22/1998 9:27:00 AM
From: Frank A. Coluccio  Read Replies (1) of 3178
 
Stephen,

>>I had no idea that much loss would occur; <<

Out of the roughly 45 Mbps a T3 has to offer, the normal overhead is like I stated, about 10 Mbps sucked up by ATM headers and management data. Another 6 Mbps gets lost to the world of SONET overhead between the 45 Mbps rate of the T3 and the ~51 Mbps OC-1/STS-1 rate. Mostly line, path and section overhead.

That's a total of 16 Mbps of overhead for every T3 ATM stream that takes SONET, enough for a 16 Mbps Token Ring to get through. In all fairness though, the latter 6 Meg also gets lost to SONET for IP T3s as well. Can't get away from it on SONET, [without kludging into the overhead space to use unused fields of data... which has actually been done in a pinch, such as using the carrier's internal data comm channels reserved for system management. But this is uncommon, and not orthodox... just thought I'd throw it in.]

But this brings to the table another set of arguments, those being that bandwidth pricing is going down and that the WAN is now taking on the attributes of a LAN... therefore no need to worry about such lavishness.

Treating bandwidth lavishly, due to its one-time sunk costs in buildings and campuses, after all, is how LANs are allowed to proliferate at the speeds that they do.

In actuality, it's difficult to draw a conclusion as to whether or not this ATM overhead is more detrimental than the much lesser amounts of superficial overhead lost in IP streams. The reason for this is that up until now, ATM has been deterministic (meaning that if it enters the pipe it almost always gets to where it is intended to go) versus the "probabilistic" nature of IP, and its sometimes circuitous (way too many hops at times) means of getting where it is intended to go, "most of the times."

Newer routing schemes are supposed to take care of this shortcoming, but the closer they get to accomplishing these kinds of improvements, the more they take on the attributes of ATM.

And ATM is getting to look more like IP all the time, too. I hear wedding bells in the not-too-distant future.... although this may be one of those open marriages, of sorts.

Regards, Frank Coluccio
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext