Plusnet Service

Customers are reporting extremely slow access to certain sites, sites listed below are some of the examples presented to us. Access to other sites such as the BBC and others are not been reported.

www.itv.com
www.myspace.com
www.ebay.com
www.facebook.com
www.amazon.com

Dial test has been performed using a customers account and this has been replicated, further reports have been raised by customers and added as examples to this problem.

If you were logged in to the portal, you could associate your username with this problem.

Problem has been passed to a network engineer for further investigation.

There appears to be some difficulty contacting Akamai to download the images on various site's homepages.

The Akamai homepage is reporting a 79% increase in the volume of daily traffic - http://www.akamai.com/

It could be that our nearest Akamai cluster is/was having problems? It may also be that Akamai have been subjected to some sort of malicious attack.

Our internal graphing suggests that something untoward happened at around 11:30pm last night.

Sites like wwwimages.adobe.com were resolving to our nearest Akamai cluster (the one we suspected problems with). This was resolving to an 84.53.137.X address earlier this afternoon.

When we looked again a little later though traffic was resolving to a 213.155.154.X address.

Our DNS is still pointing some traffic to the 84.53.137.X. address. A top-down DNS trace though is giving different IPs.

We've attempted to make contact with Akamai but are yet to hear back from them.

Investigations continue.

Our engineers are liaising with Akamai support staff to resolve this issue.

It has been noted that customers of other ISPs are seeing the same issues with Akamai.

As we've already been in contact with their support staff and provided them with as much information as possible, and we can't see anything else we can do to resolve the issue, we shall wait to see what Akamai have to say ...

Although we've not yet heard back from Akamai, our own testing, as well as that of our users, suggests that the the issue now seems resolved. As such we're downgrading this issue from a P1.