hi sebastien

your input that they reported that all the problems are solved now can be true 
or cannot be true - it´s marketing - you talked imho
with a sales guy and not with the peering engineer who really knows about the 
peering situation with other european tier-1 and tier-2  networks....

make your own picture - start your browser - define 30 or 40 targets all over 
Europe - from east to west - from north
to south, and compare one time rtt/traceroute times from cogent with other 
typical tier-1/tier-2 carriers on base of their looking glasses.
(no endless ping - just a manual ping or traceroute - to have better results 
you can repeat each path 3 times and take always the best value of all tries)
use paris or alternatively frankfurt as source for all carriers - regarding 
targets i´d recommend the webservers of the
European internet exchanges - here you can mostly be sure that it's there where 
you believe the server to be...
eg www.vix.at<http://www.vix.at> is for sure in vienna, 
www.inex.ie<http://www.inex.ie> is for sure in dublin, 
www.linx.net<http://www.linx.net> is for sure in London etc...

- then you know if they solved it. watch carefully on tracerouts via new-york - 
if you don´t see any they really made it...
I do this test for about 5 years now and repeat it every 6 month - I have a 
long list of blacklisted ip transit carriers because
of that and also knows who is getting better and who is getting worse...

think also the other way round - taking a probably bad carrier brings you bad 
inner-european routes via usa or other continents - this is known well...
but also think on the other side - if you have a larger customer base it´s also 
a problem for you, if your customers
have much communication with servers or users in europe who has this fictive 
bad carrier as their only upstream.
your customer will shout on you - no matter if you took the expensive quality 
carrier - it´s your problem - you have to solve it...
here a suggestion can be a paid peering beside the main "quality" upstream - so 
you avoid possibly bad transit ways or bad peerings to other tier-1 via this 
fictive bad carrier - but you keep
ultimately the closest and fastest way to all servers and user of this fictive 
bad carrier... so your issue is solved

a third commercial aspect is the multihoming szenario - if you pay 1 euro for 
cogent and sell for example for 5 euro to your customer -
a paid peering to cogent would be a commercial win if you deliver traffic to 
multihomed customers - as you are rated as cogent customer
you will always win the traffic path against any other big carrier who "only" 
run´s a peering with cogent because a customer is anywhere
in the world better rated as any peering. so a paid peering could commercial 
bring in much more money as it costs and solves for sure the technical part 
written above...

best regards from austria - (and sorry for english - I don´t speak french but I 
listen via google translation :)  )

Von: owner-fr...@frnog.org [mailto:owner-fr...@frnog.org] Im Auftrag von 
Gesendet: Mittwoch, 22. Juni 2011 21:35
An: frnog@FRnOG.org
Betreff: [FRnOG] Cogent


Aujourd'hui j'ai longuement discuté avec une personne travaillant chez Cogent. 
J'ai évoqué les différents problèmes lié a l'utilisation de leur réseau (ping 
élevé, pour joindre Paris de Paris je fais un détour vers Londres,  etc...)

Donc à première vue tous ces problèmes ( ?!?) seraient réglés m'a ton dit !!
En gros ils auraient enfin réussi a faire upgrader FT pour éviter les 
engorgements et problème d'accès connu récemment...

Maintenant, Cogent aurait retrouvé une stabilité et qualité de transit IP !

Avez-vous remarqués depuis ces derniers mois, une amélioration et/ou qualité de 
transit ? Perte de paquets diminué, ping normal  ...

Le mail n'a pas pour but de lancer une polémique ou troll sur Cogent (Pas 
encore vendredi ;) ), mais juste a avoir des remontées d'informations !

J'aimerais bien connaitre votre point de vue !!

Merci par avance pour vos remarques constructives.

Répondre à