Comments for page "Distributed Computing: The Survey Pattern"

(no title)

Tue, 20 Aug 2013 16:30:36 +0000

Ok, so if I don't know the locations of my peers, I'll need to discover through some other mechanism. But after that, I get a nice pattern for consensus, for example.

(no title)

Tue, 20 Aug 2013 13:12:49 +0000

You establish the topology explicitely.

(no title)

Tue, 20 Aug 2013 13:10:13 +0000

What's the underlying protocol used in discovery?

Multicast?

(no title)

Tue, 05 Jun 2012 08:05:43 +0000

Oops! Thanks for spotting it. It's fixed now.

(no title)

Tue, 05 Jun 2012 03:43:03 +0000

I think you accidentally flipped your word choice here:

Second, timeout for the survey is set to 1000 milliseconds (1 second). This is more than sufficient on a local area network, where latencies are measured in microseconds. However, on wide area network, especially if low latency (satellite or similar) links are involved you may wish to make the timeout longer.

I'm all but certain you mean it to read:

Second, timeout for the survey is set to 1000 milliseconds (1 second). This is more than sufficient on a local area network, where latencies are measured in microseconds. However, on wide area network, especially if high latency (satellite or similar) links are involved you may wish to make the timeout longer.