Hi,

Does anyone else go thorugh ECUI ?

My servers down, as is their own - does anyone know their contact phone no. or has anyone already been in touch with them?

Yes, my site is down, my email is down, their site is down ... it's been like this for at least the last 4 hours, grr. The phone no. on their whois record is 407 445 6352, but I haven't tried calling yet.

I have noticed over the last week that emails sent to me have been delayed by several hours and sometimes the mail server is briefly inaccessible. On their site they said they were upgrading to IPv6 last week, so I assumed it was because of this. But things have gone from bad to worse.

Anyone recommend a good Windows hosting co??? :bawling:

Thanks

I called them earlier.

One of their connectivity providers was in their datacentre working on upgrading their connection and managed to chop the cable belonging to another connectivity provider. These are both well-respected global backbone companies...

When I called fibre was being busily spliced. Guess it's taking longer than expected.

Originally posted by wallaby


One of their connectivity providers was in their datacentre working on upgrading their connection and managed to chop the cable belonging to another connectivity provider. These are both well-respected global backbone companies...


<Wry laughter>

Thanks for letting us know, Wallaby. I guessed there must be a mahor problem and that they were working on it, since their own site was down. I have just received a flood of email so it looks like it's back up again. I should say that I've been hosting with ECUI since January and up till now I've been very happy with their service. Hey, these things happen ...

Yup, mines back up now too.

Bit of a worry though - one of the reasons I signed up with them was ( and I quote ) :

"To help insure that your site is always online, we also have a redundant backup of DS-3 connections as well."

hmm . . . that being the case how could having a cable to a single supplier cut leave them with nothing ?

I've only been with them a few weeks and until now they've been class, have to give them the benefit of the doubt this time I think.

Will be interested to see if the reply I get on my support ticket tallies with what you were told over the phone.

and thats why you never cut the red wire.

Have had a response on the report ticket already and they did indeed confirm that the cable had been cut.

The guilty party was a Sprint engineer who cut through a C&W connection, he must have been about as popular as a turd in a swimming pool :D

They also had some serious problems earlier this year when a router went down and was not repaired for several days. Sites were available, but *very* slow. However, to their credit they kept us informed, and refunded the month's hosting fees for all sites affected. They also improved their redundancy after that to stop this kind of problem occurring (didn't allow sufficient redundancy for clumsy engineers though <g>).

Some techie person please correct me, but after working in this business for a while it appears to me that "multiple connections" does not mean all that much in terms of resilience.

Reason? Your IP addresses tie you to a particular connection/provider, which means if that connection goes, those IP addresses go, and nothing works.

Do a traceroute on a bunch of IP addresses and the crucial last few hops always end up on the same backbone provider for the same IP address, no matter where you do the traceroute from.

It depends on who's IP addys they are. Most hosts have their own, so they don't need to use an ISP's IP block.

Some techie person please correct me, but after working in this business for a while it appears to me that "multiple connections" does not mean all that much in terms of resilience. Reason? Your IP addresses tie you to a particular connection/provider, which means if that connection goes, those IP addresses go, and nothing works.
Not really. Say you have a block of IP addresses from AT&T and you also have a connection with UUNet. If you are running BGP routing on both, you are able to able to accept that IP address via either provider. They both will route in to the switch/router mesh. If AT&T goes down, you can still route in that IP via UUNet.

Now, this is a very, very simple explanation but you can accept requests for an IP address across multiple providers..

Mike,

I've seen this said before but have never seen it in practice. I've now seen a number of situations where just one provider's connection is lost but IPs from that provider are not being routed in via other providers, despite BGP4 being run. I'd like it to happen like that, but it don't seem to in the cases I have witnessed.

Have you actually had this work in practice at RackMy?

There are different variables & provider preferences that can be set-up in BGP routing but we have a NO provider preference path which allows visitors to hit the datacenter with the fastest route.

Here are some routes from different locations to show what I mean:

From Cobalt Racks:
traceroute to rackmy.com (12.25.233.110), 50 hops max, 38 byte packets[list=1]
fe0.core0.visuallink.com (208.155.64.2) 0.706 ms 0.537 ms 0.501 ms
500.Serial5-10.GW1.DCA8.ALTER.NET (65.195.226.65) 3.845 ms 2.598 ms 4.105 ms
0.so-2-0-0.XL1.DCA8.ALTER.NET (146.188.162.178) 3.540 ms 3.091 ms 3.273 ms
0.so-0-0-0.TL1.DCA6.ALTER.NET (152.63.38.69) 3.748 ms 5.190 ms 4.770 ms
0.so-4-0-0.TL1.STL3.ALTER.NET (152.63.0.41) 29.587 ms 29.933 ms 29.487 ms
0.so-3-0-0.XR1.STL3.ALTER.NET (152.63.88.242) 29.860 ms 29.792 ms 29.477 ms
185.ATM6-0.GW2.STL3.ALTER.NET (152.63.90.113) 29.735 ms 29.499 ms 29.278 ms
inflow-stlouis-gw.customer.alter.net (157.130.98.170) 80.230 ms 45.565 ms 29.772 ms
12.25.232.11 (12.25.232.11) 29.692 ms 30.539 ms 29.922 ms
rackmy.com (12.25.233.110) 29.550 ms 29.559 ms 30.824 ms
[/list=1]Now from DataPipe:
traceroute to rackmy.com (12.25.233.110): 1-30 hops, 38 byte packets[list=1]
aggr2.fe6-6.ewr.datapipe.net (64.27.64.217) 0.368 ms 0.236 ms 0.233 ms
core1.ewr.datapipe.net (216.109.128.49) 0.362 ms (ttl=253!) 0.354 ms (ttl=253!) 0.389 ms (ttl=253!)
datapipe-gw.ip.att.net (12.119.140.73) 0.731 ms 0.401 ms 0.321 ms
gar3-p354.n54ny.ip.att.net (12.123.214.165) 2.76 ms 2.88 ms 2.70 ms
gbr6-p90.n54ny.ip.att.net (12.123.1.190) 2.50 ms 2.35 ms 3.46 ms
gbr3-p90.n54ny.ip.att.net (12.122.5.114) 2.94 ms 2.22 ms 2.96 ms
gbr3-p10.cgcil.ip.att.net (12.122.2.2) 19.4 ms 20.3 ms 20.1 ms
gbr4-p30.sl9mo.ip.att.net (12.122.2.206) 25.5 ms 26.3 ms 25.7 ms
gbr6-p70.sl9mo.ip.att.net (12.122.5.189) 25.6 ms 25.7 ms 25.5 ms
gar2-p370.sl9mo.ip.att.net (12.123.24.237) 26.1 ms 25.5 ms 25.4 ms
12.125.74.10 (12.125.74.10) 29.6 ms 25.4 ms 26.1 ms
12.25.232.11 (12.25.232.11) 25.9 ms 25.8 ms 26.3 ms
rackmy.com (12.25.233.110) 27.0 ms 26.0 ms 25.3 ms
[/list=1]

If you see, they all come in to 12.25.232.11 with is the redundant Cisco Router/Switch (running HSRP which means that one router could die and the other one picks up the traffic instantly) setup at the datacenter. If I could find a path in w/ Genuity, it would show the same thing.

With this set-up, you can reach all the way to the datacenter equipment via UUNet, AT&T or Genuity. If two fail, all will route to the third provider.

This is a true redundant set-up. Hope that helps!

Hmmm. I see.

Who is Inflow -- the owners of the NOC you use?

Does my example make sense, did it help? Yes, Inflow is our datacenter partner.

Mike,

Yes, the example made sense. It's actually the first time I've seen this happening in reality. I did some checks from http://www.tracert.com/cgi-bin/trace.pl and indeed stuff does come in on various connections.

It's impressive, I have to say.

By the way, you ought to be more upfront about Inflow -- it looks like a good setup and it makes me (and others I'd guess) even more inclined to check you out for future servers, knowing you have that kind of datacentre backup.

Glad it helped :)

I would be interested to know if any of the Ecui customers that have posted on this thread have noticed a performance decrease since i.p. v6 and the cut cable incident?

Padraic: nope, all as good as before (we get good speeds from UK). Have you?

Still good for me too (excellent in fact). I had a couple of customers complain together so it made me wonder..........