As we move into the second hour without Twitter, those of us who have grown to rely on it are wondering what is going on. Some colleagues report an ongoing DOS attack on Twitter this morning [DOS = denial of service]. Appears to be easy to take down single site based services... is Facebook next?
Twitter, Facebook, GMail and all other single site based services play with the Betweenness Paradox -- ultimate power/control when in operation, ultimate fail when not. Single point of failure. Brittleness, not resiliency.
How many points of failure do you see in the diagram above? A point of failure is when one node goes down and disconnects significant parts of the network from each other. More info on Betweenness Centrality in networks.
That giant sucking sound? That's the "trust for Twitter" quickly escaping into outer space. Twitter and cloud computing in general... first it was GMail, now Twitter, what will be "strike three"?
Posts
2 comments:
Thanks for the quick analysis of Twitter outage! I also found that Network Analysis you linked to to be a nice succinct summary.
Curious to know more why you see the outage as demonstrating the "Betweeness" concept especially. Seems like the problem was the whole network was wiped out; or that it shows our network connections are brittle if they are only one diminensional--especially if that is a virtual dimension.
How appropriate this should happen the very day before you'll be speaking at a Twitter conference! It'll surely give you something new to chew on, Valdis. Hoping to perhaps see you at least in the bar afterward.
Post a Comment