by Readers

Letters to the editor: “NGN: Skeptics view IMS as way for carriers to control users””

Opinion
Oct 24, 20055 mins

Also, overstating overhead,WMM,more IT truths,CIOs and IT skills

IMS concerns

Regarding “NGN: Skeptics view IMS as way for carriers to control users”: These are valid concerns. An IMS that only replicates today’s “Soviet style,” centrally planned services model is a cost reduction for operators, not new value for users. In my opinion, if IMS is to really add value to the world, the best-of-breed IMS provider must in fact be a generic Virtual Network Enabler (VNE), allowing enterprises and other communities to have their own fixed and/or mobile virtual network operations (VNO). Such VNOs are going to happen regardless of operator support — the only questions are at what scale and volume, and who will serve and profit from them.

John Landau

Via e-mail

Overstating overhead

In Johna Till Johnson’s column, “A revised look at IPv6”, I think she is still overstating the added overhead issue.  Keep in mind that traffic flow on the Internet is, by nature, asymmetrical.  Users are mostly “downloading” data, thus the reason why the “A” in ADSL provides a workable compromise technology.  Large chunks of data are delivered in large packets and ACKed in 40-byte packets.  The place where the proportional increase is greatest will be in the “upstream” direction where it matters least.

To suggest that most packets are 40 bytes is a little misleading.  According to the sprint data 26.78% of packets are 40-byte packets (thus ACKs), which means the rest of the traffic is composed of packets of greater size than this. Each 1,500-byte packet is “worth” 37 times more bandwidth consumption than a 40-byte packet.  Counting packets instead of total consumption is not really fair.

According to the sprint data, 26.78% of packets are 40-byte and 6.28% are 1,500-byte.  There is still nearly 10 times as much data being carried in 1,500-byte packets than 40-byte packets, let alone other size packets.

While some of Johnson’s points regarding the delay in adoption of IPv6 seem reasonable, I don’t think the overhead issue is much to worry about.

Russell Skingsley

Chief Network Consultant

FPT Information System Center No 6

FPT Information System Company

Hanoi, Vietnam

Keep it simple

Regarding “WMM addresses quality of service”: Remember the strengths of the IP-stupid network vs. the legacy telco circuit-switched smart network. If your application has latency problems, isn’t it better to re-write (fix) your application, rather than re-write (fix) your network?

Seems to me that QoS adds a lot of complexity. The IP revolution is founded on KISS — Keep it simple, stupid.

Remember, the smart telco network gave us 128-parameter modems, and we had to wait for Hayes to set up a “standard 128 defaults” for modems. Let us not go back there.

Brandon Fouts

Senior systems engineer

Puget Sound Regional Council

Seattle

More IT truths

Mark Gibbs’ BackSpin column, “The truth about IT”: I’m the technology coordinator for a K-12 school district (just over 1,100 students and 120 staff). I feel there are a couple of truths of IT that got left out, based on what I see in my own little world.

Truth #7:  Users have unrealistic expectations of IT.  All technology is expected to have 100% uptime, run at blazing speeds and do what the user wants it to do, not what the user told it to do.  Going along with these unrealistic expectations are users who don’t know what they have or what it does, but insist on needing better equipment — even though they don’t use their current equipment to its potential.

Truth #8:  Everything wrong is IT’s fault.  Lights flickered?  IT’s fault.  Program had unexpected results?  IT’s fault.  Coffee tastes too weak?  The IT guy must have done something to that, too.

In some small part, I disagree with Gibbs’ Truth #5 — I don’t think security is a pipe dream.  It’s just that the more secure we make something, the more determined someone else is to hack it.  And, of course, you can’t forget the users that will always figure out a way of goofing up your security too — although usually it’s by doing something with unintended results.

Jason Brabander

Hutchinson, Kan.

Hallelujah! Mark Gibbs’ “The truth about IT” states what I have said for years. If I were collecting a salary and only working 30 hours a week, it would be called larceny. But if I work 50 hours for that same salary, I’m a “company man”! Being forced to go the “extra mile” is nothing more than institutional larceny.

David M. Babineau

West Boylston, Mass.

One of Mark Gibbs’ “truths about IT” is that security is a pipe dream. I would amend that to say perfect security is a pipe dream, in IT as well as other areas of life. We want computers to be perfect while we are willing to accept that the risk of being killed on the freeway is very real. We can manage and reduce risk, but not eliminate it. Security itself becomes a problem if we try to remove all risk — like what some people try to do with air travel — rather than manage and reduce it.

Gerald Edgar

Renton, Wash.

Good news

Regarding “U.S. CIOs see increasing demand for IT skills, says study”: This is the best news to come across my desk in a very long time. There has been an absolute deplorable lowering in generally acceptable standards for the overall IT industry employee from the perspective of more and more employers being able to dictate the market. I hope that this will now start to change.

Janita Barfield

COO

GoIP Global

New York