I've been running a few 8 hour server tests on sites hosted by Hostrocket, Future Quest and Tera Byte. These are the results of these tests from yesterday - quite interesting.
Where the Hostrocket servers down for a couple of hours yesterday? I'm rather concerned about these results, as I was almost plumping for them.

Results:

Hosted by Future Quest:
http://scp.netmechanic.com/reports/291153213027238.htm
http://scp.netmechanic.com/reports/291153235027535.htm
http://scp.netmechanic.com/reports/291153346028435.htm

Hosted by Tera-Byte
http://scp.netmechanic.com/reports/291153411028762.htm

Hosted by Hostrocket
http://scp.netmechanic.com/reports/291153433029046.htm
http://scp.netmechanic.com/reports/291153458029375.htm
http://scp.netmechanic.com/reports/291153610030248.htm
http://scp.netmechanic.com/reports/291153728031207.htm

Are these results flawed in any way?

What are your thoughts?

Host Rocket is on Alabanza, so if hostrocket was having problems on that day, then more than likely it was Alabanza.

Yes, it seems Alabanza has been having more downtime problems - some of the sites I know of on their servers were down when I attempted to access them on some different occasions.

Looks like we can definitely send that 'rock-solid' reputation right down the drain....

BC,
Do you know if VDI based servers and the VDI NOC as such are more rock solid than Alabanza? And do you by any chance know when VDI is coming up with the automated billing similar to Alabanza?

Wing.

Originally posted by TheWingThing
And do you by any chance know when VDI is coming up with the automated billing similar to Alabanza?

I believe Site5, Ultraspeed, and VDI are all trying to get this going. They probably should have pooled their efforts? I think that by the end of the year all of them will have something...

I have also done some tests the first one shows why I wish to get my wifes art page off @home's free service

http://scp.netmechanic.com/reports/291164124009018.htm

The next are for tera-byte servers

http://scp.netmechanic.com/reports/290225301026036.htm
above is for 216.234.161.13 (redirection.net) uses tera-byte

http://scp.netmechanic.com/reports/290215407019884.htm
http://scp.netmechanic.com/reports/291164007008032.htm
http://scp.netmechanic.com/reports/291164041008464.htm
http://scp.netmechanic.com/reports/291184022022308.htm

I find it difficult to really figure out just what this means. I have been pinging and getting netmechanic tests on tera-byte for over 2 months now on and off and the host ping always ranks up about 100ms on average. This seems higher than most others I have pinged. The other parts of the test, which I only get from doing netmechanic testing, I have done about 10 over the 2 months. These seem to show that dns lookup and download time are about middle of the road to slightly less than others. But Tera-byte have much more economical packages for the small guy.

The way I read these test and I may be totally off base and please tell me if I am (Steve from Tera-Byte if you read please give your thoughts) is that to get to Tera-byte servers seems to take longer than normal, when I look at the route through Neotrace it seems to go through more nodes than others that I test don't know whether this is why. Once connected though the download time seems reasonable. To me this seems fair if it takes a little longer for the opening page to happen as long as when they are connected new pages are coming down at reasonable or better speed.

Please lets have some guidance from those that know.

NetMechanic is fine for evaluating server stability, but I wouldn't rely on their results for a comprehensive evaluation of somebody's performance. What most people don't realize is that they're only checking performance from one address on one provider - so if the connection between you and them happens to be less than optimal, you'll show up as slow.

Hi,

We are aware that last night there was some connectivity problems with our server at alabanza. I think it was only affecting one of the backbones because I only heard reports of intermittant downtime lastnight from maybe 5 people. When all of the sites are unaccessible I usually get a bajillion emails.

Does the server seem slow to anyone when they physically connect to it or no? As for VDI servers and billing, we have a few servers there now, and will have a billing solution available soon to offer on dedicated servers (is that advertising? I hope not)


-Brendan
brendan@hostrocket.com

I couldnt help but notice something odd about those results. Its weird that it would get an average ping time of 11,000ms+ on one site, and 7ms on another site, when they are both on the same server? Id have to think that something is screwy with the forumla that figures that out. Im not defending our downtime or anything, but those results dont make much sense.

-Brendan
brendan@hostrocket.com

TWT,

Alabanza used to be 'the one' for rock solid NOCs, servers, and what not, but that has recently fallen by the wayside. VDI used to have only one proper line run into the NOCs, but since then they've had more lines running in and as a result, their uptime/reliability is starting to come close to matching Alabanza.

Brendan,

I do agree that some of the Netmechanic figures/formulae can be a bit screwy at times, though 95% of the time it does give a reasonably decent overall picture.... Hmmm....

ping times can help to determine problem area's but they are not the be all and end all gospel truth to connectivity.
pings are based on icmp traffic some hosts block that sort of traffic completely in order to prevent attacks, other hosts PVC off a smaller portion of pipe for them. yes I am in canada (our hosting is here as well) but that doesnt mean its a problem to host with someone farther away on the ping and traceroute reports. what really matters is connectivity how fast do the files themselves move? does the host have multiple backbone connections and enough speed for all their hosted sites? how do they react to downtime? these i think are more important questions. so essentially gthorley the reason i show up as more hops is im farther away from the starting position.

Steve
Tera-byte

I know that VDI blocks all ICMP traffic at the routers. You might not have much luck pinging us :(

Matt
mlightner@site5.com

I finally bit the bullet picked Tera-Byte and signed up yesterday.


Originally posted by keeg
yes I am in Canada (our hosting is here as well) but that doesn't mean its a problem to host with someone farther away on the ping and traceroute reports. what really matters is connectivity how fast do the files themselves move? does the host have multiple backbone connections and enough speed for all their hosted sites? how do they react to downtime? these I think are more important questions. so essentially gthorley the reason I show up as more hops is i'm farther away from the starting position.

Steve
Tera-byte

I don't understand more hops as when I do a trace with Neotrace to locations farther away some hit less nodes.

By the way did you do something in the past several days to increase connection speed because my pings yesterday and today for the sites I have monitored plus my new connection are running at less than 70ms on average whereas before they were always around the 100ms average. Maybe Shaw cable has done something to speed things up at my end.

Canadians!!! - Buy Canadian and Americans wanting the best value for their dollar should to.