Internap

[thab]

I guess I just have one question..... what on earth were they thinking with the downgrade? Is there any more info on this? (Please point me to news if I somehow missed it. I'm old and blind).

Thanks.

[Edge100x]

Are you thinking of the event we posted in the control panel?

Internap hasn't discussed their reasons publicly, but providers are added and removed from their network for a variety of reasons. Internap does not factor DDoS mitigation potential into that reasoning, which I feel is a serious mistake. DDoS attacks are more and more common in general, and quite impactful to networks, so they need to be taken seriously.

[TacTicToe]

Yea this was just stupid. Our Atlanta box is STILL unreachable. I cant connect to the RDP or anything.

What did they do as a replacement for GTT?

[kraze]

They have not replaced GTT with anyone, yet. Though, we're bringing GTT in ourselves in our Chicago center. That connection should be up and going within a few weeks. Atlanta is a different story and was due to a routing error on their part, which we're pushing them to fix quickly.

[TacTicToe]

Update @ 10:59pm EST: We continue to wait for Internap to fix their network. The problems have also made the location more susceptible to DDoS attacks, so we have seeing intermittent packet loss problems from customers being hit over the last hour.

Not happy. Even though we have multiple IPs on our box, this is a huge issue because the base IP starting with 192 is what the box seems to be tied to. Some of our servers trying to call home to master servers are failing. Web browser does not work. I can't update servers with steamcmd because it fails to resolve host. A couple of our servers do not have IP binding yet, so they are tied to the 192 IP and show as offline.

So frustrating.

[kraze]

Yes, we're definitely NOT happy either! We continue to push them to fix this problem.

As for your machine, if need be you may consider removing the IP from your adapter so you can get up and going.

[Edge100x]

The Atlanta problem today was a major screw-up by Internap. They made a sudden and significant network change in the middle of the day, and they made many mistakes along with that change that caused several big problems.

Maintenance windows are a thing for good reason. Any major change needs to be scheduled in advance, with notification, and given an appropriate window in an off-peak time of day such as the very early morning. They skipped all of that here and we suffered because of it.

These Internap missteps over the last two weeks, combined with the fact that they also steadfastly refuse to upgrade undersized links in order to help their customers with attacks, make it much more likely that we will move significant traffic away from Internap's network and physically move servers out of Internap facilities.

[thab]

Never a dull moment. Thanks for the info!

[noname]

These Internap missteps over the last two weeks, combined with the fact that they also steadfastly refuse to upgrade undersized links in order to help their customers with attacks, make it much more likely that we will move significant traffic away from Internap's network and physically move servers out of Internap facilities.

Good, especially in Europe for your Frankfurt location, better providers should be easily available.

[TacTicToe]

The Atlanta problem today was a major screw-up by Internap. They made a sudden and significant network change in the middle of the day, and they made many mistakes along with that change that caused several big problems.

These Internap missteps over the last two weeks, combined with the fact that they also steadfastly refuse to upgrade undersized links in order to help their customers with attacks, make it much more likely that we will move significant traffic away from Internap's network and physically move servers out of Internap facilities.

Yea the Atlanta issues was a huge headache for me. My chat blew up.

You've boasted here on the site and front page about how great InterNAP is and that they were the go to place for quality bandwidth. Are you saying now that these recent blunders are enough to make you reconsider your position on InterNAP as a whole? Are we going to be facing significant down times when you start making these changes?

[kraze]

The Atlanta problem today was a major screw-up by Internap. They made a sudden and significant network change in the middle of the day, and they made many mistakes along with that change that caused several big problems.

These Internap missteps over the last two weeks, combined with the fact that they also steadfastly refuse to upgrade undersized links in order to help their customers with attacks, make it much more likely that we will move significant traffic away from Internap's network and physically move servers out of Internap facilities.

Yea the Atlanta issues was a huge headache for me. My chat blew up.

You've boasted here on the site and front page about how great InterNAP is and that they were the go to place for quality bandwidth. Are you saying now that these recent blunders are enough to make you reconsider your position on InterNAP as a whole? Are we going to be facing significant down times when you start making these changes?

Internap still is good at what they do, and their process of connecting to multiple carriers and allowing their routing system(MIRO) to balance connections so users get the best routing is great. The biggest issue we face with Internap, is that they are struggling with decisions on how to handle DDoS attacks. For some very unknown reason they like to consider DDoS attacks as not 'real', when in fact they are very much real and only getting bigger. At a certain point steps will need to be taken from their end to be better equipped, and so far they've been extremely stubborn about making these steps. Sadly, in some locations we can only be as equipped as Internap allows us to be :/

What's been mentioned here is extremely preliminary, and nothing is set in stone. Any change that would involve downtime or a physical move will always be communicated to customers in-depth and in advance.

[Edge100x]

Kraze is correct. The major problem with Internap is that it doesn't take DDoS attacks seriously and actively refuses to perform necessary upgrades to help with them. In fact, recent downgrades have made things much worse in that regard.

I also have problems with their lack of communication, which I feel is very important.

MIRO works well and having multiple providers in the mix is still key to good performance, and Internap has those things.

[theRadAleks]

Any news/updates?

[Edge100x]

Internap has been opaque about any changes that they are considering internally.

On our end, we are continuing to explore other options. Any options requiring a physical move will take months and have a high cost associated with them for us, unfortunately.

[stickz]

Chicago seems to work better for my clients without GTT expect for the odd network spike. I've noticed a lot quicker download times as less people are being throttled on the outbound. Then again, I have unique requirements that average person doesn't. It renders Seattle web sync a horrible choice due to low throughput on 100's of files sent.

I've noticed that other hosting providers tend to use GTT pretty conservatively as well for the same reason. In some cases for web traffic, even, if this means downgrading to a lower tier provider like hurricane electric. If you don't have all your eggs on one basket, you shouldn't have to relay on a single transit provider to have good performance. This is especially true for locations like New York; which, have access to the NYIIX exchange point for additional providers.