Re: Re: Some problems in test case which comparing sql query performance between Ignite and Oracle

2016-10-20 Thread 胡永亮/Bob
Hi Warm-hearted person, Thanks for your help. According to your advice, I will query some field not using select *. I will add the index in the cache. To Vladislav Pyatkov, In previous mail, you said "you are always can copy data to another cache (with index) and drop this",

Re: Some problems in test case which comparing sql query performance between Ignite and Oracle

2016-10-20 Thread Yakov Zhdanov
Bob, can you please also tell what is your benchmark scenario? Do you have enough warm-up cycles to bring your app to some steady state? --Yakov 2016-10-20 12:40 GMT+03:00 Vladislav Pyatkov : > Hi Bob, > > One way to do SQL faster this is adding indexes. > 1) I do not

Re: Some problems in test case which comparing sql query performance between Ignite and Oracle

2016-10-20 Thread Vladislav Pyatkov
Hi Bob, One way to do SQL faster this is adding indexes. 1) I do not think what the estimation will be a lot improve without index. Because of the need to serialize, deserialize and move in network data. 2) Ignite does not create index on existing data, but you are always can copy data to

Re: Some problems in test case which comparing sql query performance between Ignite and Oracle

2016-10-20 Thread Alexey Kuznetsov
Bob, 1) Did you create index on Ignite cache? 2) Could you attach you cluster and cache configs? I also agree with Jörn Franke, that such simple query (no joins, simple "where condition") is not very suitable for benchmarking (unless all in your app you are using such queries a lot). --

Re: Some problems in test case which comparing sql query performance between Ignite and Oracle

2016-10-20 Thread Jörn Franke
You have to understand for what the database cache is good: lookups of single/few rows. This is due to the data structure of a cache. In this sense you use the cache wrongly. Aside of this I think select * is really the worst way to do professional performance evaluation of your architecture.

Some problems in test case which comparing sql query performance between Ignite and Oracle

2016-10-20 Thread 胡永亮/Bob
Hi, everyone My test environment: Ignite cluster has 8 nodes, every node has 8 cores CPU and 30G memory. Their network has 1000M speed. Oracle is deployed in the machine which has 32G memory and 8 cores CPU. My db table has 47535542 rows with 99 columns. When no index, the