tag:blogger.com,1999:blog-2889416825250254881.post5757319761292500585..comments2024-01-27T11:41:32.146+00:00Comments on Byte Rot: NoSQL Benchmarking - Redis, MongoDB, Cassandra, RavenDB and SQL Serveraliostadhttp://www.blogger.com/profile/05695786967974402749noreply@blogger.comBlogger10125tag:blogger.com,1999:blog-2889416825250254881.post-67625109369527927922016-03-19T20:48:34.860+00:002016-03-19T20:48:34.860+00:00what about the scripts that generates the collecti...what about the scripts that generates the collections in mongoDBYUSUF RUFAIhttps://www.blogger.com/profile/14063725693556144809noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-40263612420746886152013-02-11T20:32:43.979+00:002013-02-11T20:32:43.979+00:00Interesting article but I feel that you probably h...Interesting article but I feel that you probably hampered sql server insert speed because your choice of clustered index. Non sequential keys suchas a guid will introduce page splits slowing each insert. <br />This will in turn slow the reads as you have more datapages to traverse. A combination of sequential PK and sensible fill factor should improve the situation for sql server.dfhttps://www.blogger.com/profile/01628372863645744796noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-10892413293308747342013-01-07T23:34:57.814+00:002013-01-07T23:34:57.814+00:00Thanks for this post! I'd love to see an updat...Thanks for this post! I'd love to see an update of this post with the new RavenDB 2.0 and their BulkInsert feature.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-37472993518079044952012-11-14T10:12:34.689+00:002012-11-14T10:12:34.689+00:00The most *significant* difference between redis on...The most *significant* difference between redis on linux vs windows is the persistence and sync performance - basically the missing "fork". I don't think the test here is looking at persistence performance. While I agree that there may be slight differences in terms of standard running performance (network stack, memory allocation, etc), I wouldn't be *too* concerned by those factors.Marc Gravellhttps://www.blogger.com/profile/01023334706549710089noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-4156566291152410942012-11-14T10:03:35.476+00:002012-11-14T10:03:35.476+00:00I guess what I'm saying (above comment) is tha...I guess what I'm saying (above comment) is that I think your last-but-one paragraph needs a tweak: I think the credit there goes to Salvatore (et al), not the specific client.Marc Gravellhttps://www.blogger.com/profile/01023334706549710089noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-5459459029766646982012-11-14T09:59:40.676+00:002012-11-14T09:59:40.676+00:00For the redis performance when measuring the "...For the redis performance when measuring the "get", it isn't actually the client making it fast - I say this because you are running things sequentially and waiting for the result (via .Result) - which means it isn't even *starting* to show off the pipeline / multiplexer performance. The time you are seeing is simply: redis is really, really fast.<br /><br />The time that BookSleeve's async design really shines is when you are either:<br /><br />a: multiplexing - for example, issuing concurrent commands on a single connection from multiple threads; for reference, at StackOverflow / StackExchange we have a single BookSleeve connection per AppDomain - and it keeps up just fine<br />b: pipelining - i.e. issuing lots of commands *without* checking for the result of each before issuing the next - and then gathering the results afterwards (i.e. issue 40 "get" before accessing the .Result of the first one)Marc Gravellhttps://www.blogger.com/profile/01023334706549710089noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-36881736502816830242012-11-11T22:34:45.589+00:002012-11-11T22:34:45.589+00:00If someone is interested to further tweaks and my ...If someone is interested to further tweaks and my private conclusions on RavenDB tests then go to the raven google group: http://bit.ly/ZaLXHndario-ghttps://www.blogger.com/profile/18081266430562237223noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-26300796823209007952012-11-11T20:05:51.552+00:002012-11-11T20:05:51.552+00:00This comment has been removed by the author.dario-ghttps://www.blogger.com/profile/18081266430562237223noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-61916661665446013062012-11-11T16:50:03.533+00:002012-11-11T16:50:03.533+00:00Thanks! Yes, I intended to but forgot. I will upda...Thanks! Yes, I intended to but forgot. I will update the post now.aliostadhttps://www.blogger.com/profile/05695786967974402749noreply@blogger.comtag:blogger.com,1999:blog-2889416825250254881.post-10026729950403745882012-11-11T16:37:49.887+00:002012-11-11T16:37:49.887+00:00Great post and hey from Toronto.
This would be im...Great post and hey from Toronto.<br /><br />This would be immensely helpful if you posted versions of redis, mongo and others please :)<br />I am a big supporter of mongodb , 2.2 has a lot of improvements and i love redis for what it is, but its a very basic key/value store (yes with lists, hashes) but its basic configuration is set to "eventual persistence" , default is set to 10,000 or something records after which it'll start to persist. and also redis and others should really be tested on linux , ported versions DO NOT perform the same way.<br /><br />cheersphoenixhttps://www.blogger.com/profile/05414583744611840064noreply@blogger.com