snvijaya commented on pull request #2307: URL: https://github.com/apache/hadoop/pull/2307#issuecomment-694256719
> Sneha, > What are the likely times to > > 1. negotiate a new HTTPS connection > 2. read 4MB in a single ranged GET request > 3. read less that 4MB in a single ranged GET request, e.g. 2MB. > > If there's a fixed latency for the GET irrespective of size, then small reads are very inefficient per byte, reading the whole buffer would be justifiable. > > Also: which makes for the simplest code to write. review and maintain? Let's not ignore that little detail, especially given my experience of shipping a broken implementation of this in S3AInputStream. Hi Steve, I get your inputs and agree that observations from above points can validate a better config setting for readaheadrange. Let me try to see if I can measure up the points 1-3. Request you to give me a couple of days to get back. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
