4 ms·
<strikeout>Insights</strikeout> speculations on the performance of Microsoft's big clusters (and the services that depend on them).
by Groxx 16y ago
<strikeout>Insights</strikeout> speculations on the performance of Microsoft's big clusters (and the services that depend on them).
- rbanffy 16y ago3 authors are from Microsoft. I would expect them to be able to get current data and not need to speculate that much. As for the questions raised in the blog post, they seem reasonable. The information in the PDF doesn't quite add up.
- Groxx 16y agoBut there are no insights in the article, just things like: >Why is Cosmos (the batch-based log processing system) CPU bound instead of I/O bound? Does that make sense? That's ruminating, not answering. And the link goes to the blog post, not the PDF.
- rbanffy 16y agoThe blog post can't answer anything - the author doesn't work at MS.
- Groxx 16y agoWhich begs the question: Why does the blog post exist? To expose oddities? In which case, where are the insights that the title implies it contains? "Insights" imply that they have some insight into the questions, not that they have questions. If we were meant to dig into the PDF for information, why not link to the information directly instead of through an apparently-contentless blog post? What, if any, is the value of the linked blog post, except as a 3rd-party link to the PDF?
- rbanffy 16y agoThe oddities show that either the information is incorrect or they are not doing things in a way that current internet-sized HPC folks do it. This is very interesting in itself.