Real world distributed systems are complicated pieces of machinery that focus on making many moving parts work together.
The papers were both about high performance systems and their design. The Brewer paper focuses on the more practical implementation issues of these systems, while the Freedman paper focuses on a dynamic CDN implementation.
I would have liked for the Brewer paper to go into more detail for each other the subjects that they talked about. It is very rare to see how these systems are designed, and the more detail the better. The Freedman paper, in my opinion, was a little dry compared to the Brewer paper, and could have been shorter. I did like how they were extremely thorough, it just seemed a little too academic in its tone to me (which helped make the paper dry).
For this post, I want to focus on the way that both papers talk about network saturation. In the Freedman paper, they focus on resource saturation as a tradeoff in terms of D (data) and Q (capacity or queries per a second). They phrase performance degradation as a tradeoff between these two. In the Brewer paper, he looks at performance from a more simplistic view--cutting off bandwidth after 10g. This shows that the spectrum for dealing with performance degradation can range from complicated schemes to simple ones, and still be effective.
<3
Jon
To be fair, Freedman's paper was over a 5-year period, and he was making his observations about how the CoralCDN differed, and why, and whether or not they actually should be replicated in future systems. So it doesn't really make sense to open up his giant vat of data to pull out the important data points unless he can find a trend that supports.
ReplyDeleteAnd Brewer said in the beginning of his paper that he was going to be vague, because he was defining terms and using theory more than actual practice. But I thought the Brewer paper kind of stated the obvious. It's more logic that you'll see less throughput when you've got higher repetition, and you definitely run the risk of losing data completeness when you partition.
<3,
David
And by throughput, I meant yield. Also in the first paragraph, I accidentally a direct object. Silly me.
ReplyDeleteDavid