tag:blogger.com,1999:blog-8928181820369303270.post6586496112985597895..comments2024-01-05T13:37:57.380+05:30Comments on Billions & Terabytes: Billions of Entries and Terabytes of Data - BigMemoryAnonymoushttp://www.blogger.com/profile/00021983633501930224noreply@blogger.comBlogger6125tag:blogger.com,1999:blog-8928181820369303270.post-18128175377396870392012-11-05T16:18:44.498+05:302012-11-05T16:18:44.498+05:30@Java Geek Yeah, we have a number of customers tha...@<a href="#c7650561200894902500" rel="nofollow">Java Geek</a> Yeah, we have a number of customers that need that much BigMemory. We have tested till 7.5 TB of BigMemory for them.Anonymoushttps://www.blogger.com/profile/00021983633501930224noreply@blogger.comtag:blogger.com,1999:blog-8928181820369303270.post-76505612008949025002012-11-05T15:30:36.295+05:302012-11-05T15:30:36.295+05:30did you actually work with a project which has tha...did you actually work with a project which has that much data requirement ?Java Geekhttp://java67.blogspot.jpnoreply@blogger.comtag:blogger.com,1999:blog-8928181820369303270.post-72498009611018030862012-05-30T09:53:28.670+05:302012-05-30T09:53:28.670+05:30Average latency is the average latency of all the ...Average latency is the average latency of all the transactions including read and write. <br /><br />The avg latency for reads were around 25 ms, while writes were 20 ms. <br /><br />Reads have a higher latency because we are hitting terracotta server for each entry. This can be overcome by using BigMemory at the application level too.Anonymoushttps://www.blogger.com/profile/00021983633501930224noreply@blogger.comtag:blogger.com,1999:blog-8928181820369303270.post-13088897966738517932012-05-30T09:08:11.543+05:302012-05-30T09:08:11.543+05:30Yes, we ran some tests with increasing write ratio...Yes, we ran some tests with increasing write ratio (keeping the no. of keys same) so effectively increasing no. of updates. It gives almost flat line performance. <br /><br />Are you talking about increasing the number of keys over time too?Anonymoushttps://www.blogger.com/profile/00021983633501930224noreply@blogger.comtag:blogger.com,1999:blog-8928181820369303270.post-91827592444557321362012-05-16T23:37:15.030+05:302012-05-16T23:37:15.030+05:30Can you elaborate on the avg latency? How much tim...Can you elaborate on the avg latency? How much time on avg does an individual *read* and *write* operation take once the caches are primed ? This was not clear from the blog post.PGShttps://www.blogger.com/profile/05715928735245754149noreply@blogger.comtag:blogger.com,1999:blog-8928181820369303270.post-57273206771096175652012-05-15T22:35:01.801+05:302012-05-15T22:35:01.801+05:30What would happen if you increased the writes/read...What would happen if you increased the writes/reads ratio to a more real world work load instead of 10/90?<br /><br />Can the off heap allocator keep up?Ashwin Jayaprakashhttps://www.blogger.com/profile/14487665820971997426noreply@blogger.com