[NLNOG] EU IXP pricing

Arnold Nipper arnold at nipper.de
Wed May 25 21:26:37 UTC 2016


Hi Jac,

On 25.05.2016 15:08, Jac Kloots wrote:

> 
> On Wed, 25 May 2016, Arnold Nipper wrote:
> 
>> On 25.05.2016 09:37, Florence Lavroff wrote:
>>> Beste NLNOGers,
>>> Dear Job,
>>>
>>> Thanks to you (and other contributors) for your extensive research and
>>> useful tool.
>>>
>>> I also believe it would add value to create a separate column for NRC /
>>> set-up fees.
>>>
>>> You may also want to add an extra one for the IXP business model
>>> (commercial/for profit, non-for profit, private … , as Dave did in his
>>> slide #8).
>>>
>>
>> Good point! ... what other metrics come to mind?
> 
> I Agree with the additional columns for NRC (setup fees) and/or
> Membership fees.
> 
> Any extra additional columns could be very usefull if you run an IXP.
>

Ooops ... even after having thought about it three times, I do not yet
understand. Imho it would be useful for anyone to make a more educated
decision which IXP to chose

> For the purpose Job made this overview I think that any additional
> column only hides the information that this sheet tries to compare.
> 

If the purpose is to have a one dimensional view only then I fully agree.

> Regards,
> 
> Jac
> 

Cheers, Arnold

>>
>> * Is traffic important?
>>
>> * Number of participants?
>>
>> * Are key players there?
>>
>> * How well does the IXP cover the Metropolean Area (in terms of
>> presence in most important DC)?
>>
>> * Is financial stability a matter?
>>
>> * how well does the IXP grow?
>>
>> * stability, i.e. uptime?
>>
>>
>> Met vriendelijke groeten,
>> Arnold
>>
>>
>>> Met vriendelijke groeten,
>>> Flo
>>>
>>> 2016-05-23 8:20 GMT+02:00 Andre van Vliet <andre at solidhost.com
>>> <mailto:andre at solidhost.com>>:
>>>
>>>    >> 2) NetNod MRC is door twee gedeeld omdat je per set van twee
>>> poorten
>>>    >>    connect.
>>>    >> 3) Eventuele jaarlijkse membershipfees zijn in de MRC verwerkt
>>>    >>
>>>
>>>    >Might be useful to have a separate column for this, because the
>>> table is
>>>    >easier to maintain and fees are more transparent
>>>
>>>     I agree. This would be particularly relevant in scenarios when you
>>>     bundle multiple ports under the same membership, which is quite
>>> common.
>>>
>>>     Details like that aside: very useful Job, thanks for sharing!
>>>
>>>     -André
>>>
>>>     -----Original Message-----
>>>     From: NLNOG [mailto:nlnog-bounces at nlnog.net
>>>     <mailto:nlnog-bounces at nlnog.net>] On Behalf Of Arnold Nipper
>>>     Sent: zondag 22 mei 2016 16:36
>>>     To: Job Snijders; NLNOG
>>>     Subject: Re: [NLNOG] EU IXP pricing
>>>
>>>     Beste NLNOGers,
>>>
>>>     Ik ben nieuw op de lijst en mijn gesproken Nederlands is zeer
>>>     rudimentair. Hence excuse me if I switch to English now. Way
>>> easier for
>>>     your eyes :)
>>>
>>>     On 22.05.2016 11:34, Job Snijders wrote:
>>>    > Beste NLNOGers,
>>>    >
>>>    > Naar aanleiding van Dave Temkin's GPF praatje [1] en de stemronde
>>>    > afgelopen week bij AMS-IX over de commerciële strategie (LTCS
>>>    > 2016-2019), heb ik een overzicht gemaakt van de huidige
>>> gepubliceerde
>>>    > prijzen van diverse Europese publieke internet exchanges om beter
>>>    > inzicht te krijgen in hoe de exchanges zich onderling verhouden.
>>>    >
>>>    >     EU IXP megabit/sec cost spreadsheet:
>>>    >
>>>     
>>> https://docs.google.com/spreadsheets/d/18ztPX_ysWYqEhJlf2SKQQsTNRbkwoxPSfaC6ScEZAG8/edit#gid=0
>>>
>>>    >
>>>    > Het is interessant om te zien hoe nieuwe technologie zoals 100GE
>>>    > significant de effectieve prijs per megabit drukt. Onze vaderlandse
>>>    > juwelen NL-IX en AMS-IX zijn per mbit op 100GE (in vergelijking met
>>>    > 10GE) 38% en respectievelijk 50% goedkoper.
>>>    >
>>>
>>>     really very, very useful and well done, Job!
>>>
>>>    > Enkele opmerkingen:
>>>    >
>>>    > 0) Alles non-Euro currency rekent Google Spreadsheets om naar
>>> Euro met
>>>    >    recente data.
>>>    > 1) Ik leg twee profielen: 85% utilisation ("vol") en 40% util
>>>    >    ("gemiddeld") naast elkaar.
>>>    >
>>>
>>>     Where do these figures 40% and 85% come from? From a reasearch
>>> paper I
>>>     remember that you are able to load a 10G to over 90% w/o seeing
>>> PL even
>>>     due to microbursts. And the higher the bandwidth of the link the
>>> higher
>>>     the load can go.
>>>
>>>     The 40% may come from the ratio of peak traffic seen at an IXP to
>>> the
>>>     connected bandwidth.
>>>
>>>     For comparison just between IXP's it doesn't matter which
>>> percentage you
>>>     take. Otoh if someone is looking at the prices s/he might be
>>> attempted
>>>     to compare this with transit prices (imho comparing prices at IXP
>>> with
>>>     transit is like comparing apples with oranges imho. But that's
>>> another
>>>     story)
>>>
>>>    > 2) NetNod MRC is door twee gedeeld omdat je per set van twee
>>> poorten
>>>    >    connect.
>>>    > 3) Eventuele jaarlijkse membershipfees zijn in de MRC verwerkt
>>>    >
>>>
>>>     Might be useful to have a separate column for this, because the
>>> table is
>>>     easier to maintain and fees are more transparent
>>>
>>>    > 4) Eventuele Setup/installatie fees worden over 3 jaar afgeschreven
>>>    >    en ook in de MRC verwerkt. Cabling fees zijn niet geincludeerd.
>>>
>>>     Might be useful to have a separate column for this, because the
>>> table is
>>>     easier to maintain and fees are more transparent
>>>
>>>    > 5) Enkel openbaar gepubliceerde prijslijsten zijn gebruikt,
>>> behalve in
>>>    >    het geval van INEX waar IRC informatie is gebruikt.
>>>    > 6) De gemiddelde IXP poort schijnt maar voor voor 40% gebruikt te
>>>    >    worden, dit verschilt natuurlijk van netwerk tot netwerk!
>>>    > 7) Waar product keuze is, word in het overzicht LR4 optics
>>>     gebruikt, en
>>>    >    altijd gekozen voor de meest 'unrestrictive' variant op policy
>>>    >    gebied.
>>>    >
>>>    > "Hat tip" voor Samer Abdel-Hafez, hij heeft meegeholpen met het
>>>     opsporen
>>>    > van alle prijslijsten.
>>>    >
>>>    > Met vriendelijke groeten,
>>>    >
>>>    > Job
>>>    >
>>>    > [1]: https://www.peeringforum.com/the-real-cost-of-public-ixps/
>>>    >
>>>
>>>     Met vriendelijke groeten
>>>     Arnold
>>>     --
>>>     Arnold Nipper
>>>     email: arnold at nipper.de <mailto:arnold at nipper.de>      phone: +49
>>>     6224 5593407 2 <tel:%2B49%206224%205593407%202>
>>>     mobile: +49 172 2650958 <tel:%2B49%20172%202650958>      fax:   +49
>>>     6224 5593407 9 <tel:%2B49%206224%205593407%209>
>>>
>>>     _______________________________________________
>>>     NLNOG mailing list
>>>     NLNOG at nlnog.net <mailto:NLNOG at nlnog.net>
>>>     http://mailman.nlnog.net/listinfo/nlnog
>>>
>>>
>>>
>>>
>>> -- 
>>> Florence Lavroff / lavroff at google.com <mailto:lavroff at google.com> / +31
>>> 20 504 5474 / +31 6 11 01 55 80
>>> Peering and Content Distribution / Google EMEA
>>>
>>>
>>> _______________________________________________
>>> NLNOG mailing list
>>> NLNOG at nlnog.net
>>> http://mailman.nlnog.net/listinfo/nlnog
>>>
>>
>>
>>
> 
> 
> 
> _______________________________________________
> NLNOG mailing list
> NLNOG at nlnog.net
> http://mailman.nlnog.net/listinfo/nlnog
> 


-- 
Arnold Nipper / nIPper consulting, Sandhausen, Germany
email: arnold at nipper.de      phone: +49 6224 5593407 2
mobile: +49 172 2650958      fax:   +49 6224 5593407 9

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: OpenPGP digital signature
URL: <http://mailman.nlnog.net/pipermail/nlnog/attachments/20160525/a74ac081/attachment-0001.sig>


More information about the NLNOG mailing list