=?iso-8859-1?q?Gerard=20Torin?= wrote: > > Hi guys, > > Anybody know how is built the ATM4S Bandwith?. For example, In > ATM`s link of 34Mbps, Is true that 4Mbps is just only header?, > I ask that, because actually my company has it. But we don4t > reach the maximun bandwith of 34Mbps. Yesterday, we did stress > test in the link and just only reach 30Mbps.
Bandwidth means capacity. It is a given. The amount of bandwidth you have is dictated by your provider and the type of interface you are using. It is not measured. It doesn't change (unless you have some technology that uses dynamic bandwidth allocation, such as multilink PPP). What you seem to be doing is measuring throughput. Throughput is measured. It's the amount of data per timeframe that can be sent. Because they have the same unit of measurement (bits per second or something similiar) people think bandwidth and throughput mean the same thing. They don't. Just want to encourage people to use the terms correctly! :-) Now, throughput can measure numerous different things. If you had a tool that could send raw, unframed bits as fast as the interface allows, throughput could equal bandwidth, assuming the link is not used by anything but your testing tool. Since such tools aren't very common, and, of more importance, don't resemble real-world applications, that's not how throughput is really measured. What you really care about is the user's experience. So you care about application-layer throughput. Some tools let you measure the amount of user data that is sent per unit of time. This refers to the application-layer payload. It leaves out overhead caused by headers at the data-link, network, transport, and application layer. Or you can measure throughput using a tool that does count data-link layer overhead, or network, or whatever. If you don't know what your tool is measuring, find out. Otherwise your results are meaningless. Throughput, depending on what layer you are measuring, can be affected by numerous factors: * packet header overhead (and ATM cell overhead) * errors, resulting in retransmissions and dropped frames * dropped frames at internetworking devices due to buffer overflows * media contention on shared links * protocol behavior, including the need to find a resource, set up a connection, ack data, etc. * RAM access speed at end systems and internetworking devices * hard drive access speed at end systems * processing required at end systems and internetworking devices * software inefficiencies And about a zillion other things, depending on what you are actually measuring, which is not bandwidht, but is throughput at some layer. :-) _______________________________ Priscilla Oppenheimer www.priscilla.com > > I thanks any comment. > > > --------------------------------- > Yahoo! Messenger > Nueva versión: Super Webcam, voz, caritas animadas, y > más #161;Gratis! > > Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=71999&t=71937 -------------------------------------------------- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]