dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
1684
share rss forum feed

MaynardKrebs
Heave Steve, for the good of the country
Premium
join:2009-06-17
kudos:4
reply to InvalidError

Re: Bury the Industry Minister in complaints!

said by InvalidError:

said by graniterock:

IMHO, Complaining that you don't like the decision is one thing. Being able to articulate why it was the wrong decision and what the right one should be is better.

Many good and technically valid points were made in various proceedings, sometimes well, sometimes clumsily, but almost every time they got ignored or (conveniently?) misunderstood.

For example:
1- Incumbents claim load-balancing, fail-over and link aggregation are things that require some form of advanced voodoo magic even though it is a standard feature of just about any managed Ethernet switch/router manufactured in the past 10+ years. Most of the benefits can be achieved with almost no effort by simply enabling LACP.
2- Bell claims that 10GbE AHSSPIs are not viable below 3Gbps but why should Bell care how much load is on individual AHSSPIs? The ISPs already pay a separate per-port fees for that which should cover all capacity-invariant costs of providing that link so Bell's port costs are covered even at 0Gbps.
3- Incumbents say billing for aggregate capacity or "dynamic bandwidth allocation" is too complicated but all it requires is tracking bandwidth in 1-5 minutes increments, summing graphs across all links and bill based on the total, something that can be done fairly easily with tools like MRTG and would be even easier if LACP was enabled so usage statistics would only need to be gathered on a per-LAG basis.
4- Incumbents claim they need to charge on a per-link basis to "share risk" with 3rd-party ISPs but this makes no sense since the incumbent has already incurred all possible bandwidth costs by the time traffic gets aggregated to the POI, it does not make much difference how wider/faster the interface between the POI and ISP is at that point.

This is by no means intended to be an exhaustive list but these are my top-4 issues.

On that 3rd one though, I think CNOC and friends tripped over themselves by calling it "dynamically allocated bandwidth" since what they really want is for ports to be allowed to run at max-speed and then bills be adjusted to the excess of total capacity used beyond commitment. No "dynamic allocation" here, only dynamic billing based on total traffic across all links rather than per-link basis. Much simpler.

You just wrote CNOC's R&V.
Hope you get paid for it.


AOD
Premium
join:2008-01-24
Etobicoke, ON
kudos:1
Reviews:
·TekSavvy Cable

said by MaynardKrebs:

said by InvalidError:

said by graniterock:

IMHO, Complaining that you don't like the decision is one thing. Being able to articulate why it was the wrong decision and what the right one should be is better.

Many good and technically valid points were made in various proceedings, sometimes well, sometimes clumsily, but almost every time they got ignored or (conveniently?) misunderstood.

For example:
1- Incumbents claim load-balancing, fail-over and link aggregation are things that require some form of advanced voodoo magic even though it is a standard feature of just about any managed Ethernet switch/router manufactured in the past 10+ years. Most of the benefits can be achieved with almost no effort by simply enabling LACP.
2- Bell claims that 10GbE AHSSPIs are not viable below 3Gbps but why should Bell care how much load is on individual AHSSPIs? The ISPs already pay a separate per-port fees for that which should cover all capacity-invariant costs of providing that link so Bell's port costs are covered even at 0Gbps.
3- Incumbents say billing for aggregate capacity or "dynamic bandwidth allocation" is too complicated but all it requires is tracking bandwidth in 1-5 minutes increments, summing graphs across all links and bill based on the total, something that can be done fairly easily with tools like MRTG and would be even easier if LACP was enabled so usage statistics would only need to be gathered on a per-LAG basis.
4- Incumbents claim they need to charge on a per-link basis to "share risk" with 3rd-party ISPs but this makes no sense since the incumbent has already incurred all possible bandwidth costs by the time traffic gets aggregated to the POI, it does not make much difference how wider/faster the interface between the POI and ISP is at that point.

This is by no means intended to be an exhaustive list but these are my top-4 issues.

On that 3rd one though, I think CNOC and friends tripped over themselves by calling it "dynamically allocated bandwidth" since what they really want is for ports to be allowed to run at max-speed and then bills be adjusted to the excess of total capacity used beyond commitment. No "dynamic allocation" here, only dynamic billing based on total traffic across all links rather than per-link basis. Much simpler.

You just wrote CNOC's R&V.
Hope you get paid for it.

well played.