Jeroen Wunnink schreef op 2018-04-25 11:45:
We have AMSIX, NLIX and a few other smaller fragmented exchanges already that span multiple larger+smaller datacenters and are carrier/vendor-neutral (more or less, let's not get into that ;-). Thus I'm a bit doubtful on the success of a 'local' datacenter IX that's vendor specific in a small country like NL.
In the US this makes more sense, given that everyone is at pretty much the same large hubs and there's not that many mature alternatives with a decent userbase. Competition is good, but given the maturity of some existing exchanges here and pretty much every AS being on one or multiple exchanges as-is (pun not intended), it might be a hard sell to get people on the Equinix specific ones as well.
I'm not seeing any real added value on launching these here unless I'm missing a big incentive to do so (i.e. free IX connectivity with your Cloud Exchange port or vice versa?)
I agree that I also do not see much if any value purely from the "Exchange" point of view - the biggest value of an Exchange is basically the number of suitable peers present . So I doubt many people will want to _go_ there primarily to connect to the iX . However, if you are _already_ present in this Datacenter and wish to peer with someone who is also present, using the 'in house iX' could be a nice and convenient service from the datacenter operator . Depending on in-house iX pricing, traffic volumes, resiliency wishes in your design, and (port)prices of your equipment and that of your peers the in-house iX may well be preferable to private peering , or both peers backhauling to the nearest external exchange . If the DC operator already has a sufficiently large network design/operations team, the marginal cost of operating an 'in-DC exchange service' should be quite low , and imo then it can be a nice value add [not unique selling point] for their customers. Not to mention the marketing value , as this thread shows :-) Boudewijn
Jeroen Wunnink Integration Engineering Manager
www.gtt.net <http://www.gtt.net/>
On 24/04/2018, 18:55, "NLNOG on behalf of Arnold Nipper" <nlnog-bounces@nlnog.net on behalf of arnold@nipper.de> wrote:
[ Sorry for switching to Enlish, but my Dutch is to bad ]
On 24.04.2018 13:08, niels=nlnog@bakker.net wrote: > * job@instituut.net (Job Snijders) [Tue 24 Apr 2018, 12:41 CEST]: >> Er is (alweer) een nieuwe IX in Nederland, sommigen zagen deze al van >> verre aankomen: >> >> >> https://url.emailprotection.link/?a2CpwB7v-YuNg3Hr3KRdFPHjkSLCheU7rsmqNfVTOT... >> >> >> Wat is de impact op NL-IX, AMS-IX, SPEED-IX & Asteroid dat nu een >> datacenter naast crossconnects, power en ruimte, ook IX diensten bied? >> >> Profiteren wij als operators van meer competitie in deze markt, of zal >> dit leiden tot minder competitie? Hoe erg is fragmentatie? > > Wat is het punt van een Internet exchange waar je alleen kan peeren met > andere partijen in hetzelfde datacenter? Dat kan je dan ook direct > doen. AMS-IX biedt operators juist vrijheid van datacenterkeuze. >
According to Asteroid this was/is the real intention of an IX. YMMV and mine also does ;-)
Competition is always good. However where I could see an issue is when DC operators also jump the bandwagon.
From a network pov a good choice of IXes is always good as they don't have to put all their eggs in one basket.
Let's see how this all works out.
Groeten, Arnold -- Arnold Nipper email: arnold@nipper.de mobile: +49 172 2650958
_______________________________________________ NLNOG mailing list NLNOG@nlnog.net http://mailman.nlnog.net/listinfo/nlnog