Hi justone -
If this is true, it seems to me the IP over ATM.... etc. has the better chance at the core/metro for data networks, particularly if bandwidth is free. But now I've stepped into the religious debate (war), which I tremble to do given your bold declamatory cognomen. However, it may be risky to invest in 10, 100, 1000 or 10^x Gb plays, particularly as the last mile has a lot more issues than just high bandwidth access and flexible management.
Since I've been reading this thread, I have seen the idea of IP over ATM revisited many times. My sense is that ATM switching, in a network that is a subset of a faster optical network, is an alternative of real merit. I recall posts, by MikeM, with stats on increasing use of ATM, I think WRT sales by Alcatel after they absorbed Newbridge.
The debate goes on, while ATM usage increases.
What problem(s) with ATM cause this debate?
Ethernet does not seem to offer sufficient manageability unless network usage is 'way below its theoretical capacity.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
After many discussions of QoS, it came as a surprise to me that 'the equipment needed for business-level measurements is not yet commercially available.'
The following article seems to indicate that accurate measurement of network QoS is on its way, and that such measurement will provide a means of differentiating between carriers...[emphasis added]
Network Infrastructure
True IP QoS to be possible in 2001
By Simon Marshall, Total Telecom, in Paris
11 September 2000
It will be at least another year before IP service providers will be able to measure quality of service (QoS) accurately enough to guarantee and differentiate service levels to corporate customers.
Bob Mandeville, president of packet measurement start-up CQOS, told Total Telecom on day one of the IP Infrastructure Summit in Paris on Monday that the equipment needed for business-level measurements is not yet commercially available.
Mandeville said IP service providers have problems with metrics such as one-way latency, jitter, load imbalance on routers and pathway congestion to servers. "The equipment to measure scalability in particular is unavailable," he said. "This means service providers' understanding of their own network is, to say the least, modest. When we tested them, the results we drew were almost universally bad. Yet when I ask them if there's the equipment available off-the-shelf to purpose-test those networks, the answer from them is always 'no'."
The ability to market services on the basis of effective QoS measurement would, according to Mandeville, offer service providers the opportunity to differentiate their services and more quickly recoup some of the cost of building out the network.
"The dynamics of competition between IP service providers will inevitably drive them to offer measurement as part of their service portfolios," Mandeville told Total Telecom. "CQOS has had a lot of interest from corporates that want to be able to measure QoS on their own networks - [by doing this] they could have a better network than the service provider leasing it to them," he said.
He added that accurate QoS measurement was particularly important to specific market sectors, citing finance, international travel and transport as examples. In particular, faster, more accurate IP services could provide a boost to banking and stock trading organizations. "Service providers and corporates will be looking to differentiate levels of service within a year," said Mandeville. "They will base their business models on this, but the technology to do this right now [is not yet available]. CQOS will be offering outsourced QoS measurement and management services by the first quarter of 2001," he claimed.
totaltele.com |
