Aardvark Daily

Please visit the sponsor!

The now very fragile Internet

We're told that the Internet grew out of DARPA's attempts to build a resilient network that could withstand multiple failures and keep on working.

Well, based on last night's online experience, I'd have to say that was a joke.

At about 8:30pm I found that neither my YahooMail nor my GMail accounts were responding.

A quick check of my favourite sites and some traceroutes indicated that the local interweb was working just fine but some routes to US networks were badly broken.

No problems -- I'll just use some local sites and some UK-based ones.

Ah... but these days, that's not how the Net works.

I thought I'd visit the BBC's website.

It started loading just fine -- and then stalled -- waiting on an unreachable server at ad.doubleclick.net

No worries -- some US sites (such as Aardvark and the Aardvark Forums (interestingprojects.com) were still working so I figured I might try Slashdot.org as well.

Just like the BBC's site, Slashdot started loading and then stalled -- waiting on that bloody ad.doubleclick.net again!

How about the UK DailyMail website?

Well it loaded but unfortunately the style sheet didn't and the route to one of their servers seemed to "bounce" off the USA's crippled network:

Hop	Hostname	IP	Time 1
3	mdr-fid-int.akbr5.global-gateway.net.nz	202.37.244.222	1119.142ms
4	ae4-10.akbr5.global-gateway.net.nz	202.37.244.221	30.629ms
5	ae1-2.akbr4.global-gateway.net.nz	202.50.232.77	26.302ms
6	ae1-10.tkbr9.global-gateway.net.nz	202.50.232.37	26.602ms
7	ae2-3.labr5.global-gateway.net.nz	203.96.120.142	150.548ms
8	ae0-3.lebr6.global-gateway.net.nz	203.96.120.86	151.116ms
9	xe-10-1-0.edge2.LosAngeles9.Level3.net	4.53.230.13	160.045ms
10	ae-2-70.edge1.LosAngeles9.Level3.net	4.69.144.74	151.018ms
3	mdr-fid-int.akbr5.global-gateway.net.nz	202.37.244.222	1164.054ms
11	lap-brdr-03.inet.qwest.net	63.146.27.33	151.217ms
12	no	reply	*
13	65.113.16.54	65.113.16.54	151.124ms
14	news.bbcimg.co.uk	63.238.2.250	157.630ms

What the hell's going on there.

What the events of last night seem to prove is that most websites are now a complex interleave of content from a wide range of different servers, often in different locations around the globe. When one part of that network goes down, a massive number of other sites can be immediately crippled.

Instead of ending up with the multiply fault-tolerant network that DARPA first conceived, we now have a house of cards -- where a single failure on a critical part of a network *somewhere* in the world can affect sites right across the globe.

Is this a good thing?

Is it surprising that this doesn't happen more often?

And as a footnote -- it's taken me 35 minutes to write this column and all the time I've been on hold, trying to get through to Telecom's DSL helpdesk and listening to the musak they play in your ear at such times.

Obviously a lot of folk are affected by this problem and the help desk is being swamped.

Why do you think that is?

Probably because nobody has even bothered to update the Telecom DSL network status page to advise that there are problems. As I type this at 9:00pm, the status message reads "Broadband network all clear". (Update: it was changed by 9:06pm when I checked again to "Issues Accessing Some Websites").

My gawd -- surely Telecom can be a little more "on the ball than that"?

How long does it take to identify that the helpdesk load has jumped by 1000% and simply update that message?

Please visit the sponsor!

Have your say on this...

PERMALINK to this column

Oh, and don't forget today's sci/tech news headlines

Rank This Aardvark Page

Change Font