Distributed system definition at DEC SRC bulletin board (1987)(lamport.azurewebsites.net)
lamport.azurewebsites.net
Distributed system definition at DEC SRC bulletin board (1987)
https://lamport.azurewebsites.net/pubs/distributed-system.txt
4 comments
I think this is worth looking at to see the history of why building distributed systems is hard.
Also this, "A distributed system is one in which the failure of a computer you didn't even know existed can render your own computer unusable." Is a good kind of funny explanation of what can go wrong.
Lamport obviously did spend his career fixing the problem and we owe a tremendous amount to him.
Also this, "A distributed system is one in which the failure of a computer you didn't even know existed can render your own computer unusable." Is a good kind of funny explanation of what can go wrong.
Lamport obviously did spend his career fixing the problem and we owe a tremendous amount to him.
> Not sure why this is being posted now.
This kind of thing often happens when an interesting link is mentioned as a comment to some recent event. I linked to it somewhat recently (https://news.ycombinator.com/item?id=23790516), but that was over two weeks ago, so it probably wasn't my comment which led to this being posted now.
> However, it seems out-of-touch today, because of years of work in building more reliable systems out of distributed components has made the internet and scalable internet services possible.
And yet, it still happens. That comment I posted was about a computer most people didn't even know existed, making their own phones (which can be considered computers nowadays) unusable. A more recent event would be the Garmin servers going down, making many devices unusable.
This kind of thing often happens when an interesting link is mentioned as a comment to some recent event. I linked to it somewhat recently (https://news.ycombinator.com/item?id=23790516), but that was over two weeks ago, so it probably wasn't my comment which led to this being posted now.
> However, it seems out-of-touch today, because of years of work in building more reliable systems out of distributed components has made the internet and scalable internet services possible.
And yet, it still happens. That comment I posted was about a computer most people didn't even know existed, making their own phones (which can be considered computers nowadays) unusable. A more recent event would be the Garmin servers going down, making many devices unusable.
Right around this time, the VMS side of the house was releasing VAXcluster, which was really quite robust, and ahead of its time. Lots of ideas that are now in systems like Kubernetes have very close parallels to VAXcluster systems: nodes that dynamically join the cluster, load failover to other nodes, shared resources like disk and security policies, asymmetric nodes etc.
https://homes.cs.washington.edu/~levy/vaxclusters.pdf
https://homes.cs.washington.edu/~levy/vaxclusters.pdf
However, it seems out-of-touch today, because of years of work in building more reliable systems out of distributed components has made the internet and scalable internet services possible.
It is perhaps not without irony that one of the key idea drivers in this space was Lamport himself, with early papers on clocks and synchronization, and later work on Paxos, which is found at the heart of many of the most reliable systems in the world.
Perhaps we should call this a "Lamport's Irony"; when you complain about something (jokingly) and then spend your career fixing the problem you complained about.