see my last written recommendations in the cogent thread a few more suggestions to find out by yourself if sparkle is an interesting alternative
1) see a customer base here: http://www.cidr-report.org/cgi-bin/as-report?as=AS6762&view=2.0 if you know anyone of these you can ask 2) analyze the network maps - especially the fiber and ip path´s from your point of interconnect - paris seems to be good interconnected as you see 6 different path´s in 6 different geographic directions. but be careful - it can also be a logical picture which doesn´t tell you anything about the really used fiber routes. compare this picture if you are located for example in warsaw - so each carriers doesn´t fit for everyone - it depends where you are and where you want to communicate to http://www.tisparkle.com/default.aspx?Lang=ENG&idPage=2158&initSez=1 3) analyze the peering memberships - a pop doesn't help if it is not interconnected with the national ix´es... analyze the membership lists of france-ix, panap, equinix.... etc... but also other main hubs around you - Amsterdam, London, Madrid, Milano, Zurich, Frankfurt.... especially have an eye if you see the jump-over to other large networks in Amsterdam - a, for french customers well suited carrier should be directly interconnected to most other tier-1 and near-tier-1 direct in paris and not in amsterdam, frankfurt or London - you can find out typical via looking glass or transit-trial. but think also on private interconnects - if networks are heavily interconnected on private base a peering-point membership is not a must-have. (but it´s suspicious to have no direct interconnect to medium and small networks then) 4) analyze the peering strategy - does your carrier of choice peers only with other tier-1 and with no one else or does it come down with peerings to a medium size (eG interoute or lambdanet) or even to a tiny size like he.net - expect potential longer ways to smaller networks from carriers who doesn´t peer with smaller networks - you can find out the peering strategy by analyzing the peering matrix of some internet exchanges who publish this information. 5) if you are a carrier who is focused on getting many large peerings by yourself it´s often a well know strategy to take only the two largest tier-1 carriers and nothing else - this traffic path is the worst case for many medium sized network who must pay for this traffic and then have a potential larger interest in peering with you. if you are not involved in this, a mixture of a large worldwide tier-1 with a well peered european player can potentially make more sence... so a few more suggestions for you - hopefully it´s a help bernd -----Ursprüngliche Nachricht----- Von: owner-fr...@frnog.org [mailto:owner-fr...@frnog.org] Im Auftrag von Renaud Gesendet: Donnerstag, 23. Juni 2011 09:16 An: Liste FRnoG Betreff: [FRnOG] SPARKLE Bonjour tout le monde, Bon je vois que le débat fait rage concernant COGENT !!! Pour ma part je suis actuellement en réflexion concernant SPARKLE (telecom italia) notamment pour du trafic IP plutôt hors de france. Puisque le sujet va un peu au-delà de la simple qualité mais se porte aussi du coté "opérateur sympa plutôt que voyou", quelqu'un est-il (ou a été) client chez eux ? Renaud --------------------------- Liste de diffusion du FRnOG http://www.frnog.org/ --------------------------- Liste de diffusion du FRnOG http://www.frnog.org/