YCSB Testing Mysql,mongodb,tokumx,couchbase Performance

The test is done by a colleague, and this is just a collection. Test Description: 1, the amount of data is 3kw records, 11 fields per record, a primary key, the key is a character type, similar to: user****, followed by the value of the other 10 fields for the character type, 100 characters, the record length of about 1KB2, using YCSB for stress testing. Specific types of reference to the previous documents 3, 30 million data, to the current database can be cached into the db (subsequent test cache 80%,50%,30% performance) 4, test items include: Data loading, is insert data                       YCSB standard load A, load B, load C, load D, load E, load F test                       All 64 thread-based, analysis is optimal                   &NB Sp   MySQL test environment for innodb_buffer_pool_size=20g                      m ysql_100g for innodb_buffer_pool_size=100g                      Mongod b for cachesizegb=20g                      mongodb_100g for cachesizegb=10 0G load a:read=0.5,update=0.5,insert=0,scan=0 (no range query) load b:read=0.95,update=0.05,insert=0,scan=0 (no range query) load c:read=100  ---Read-only application, scan=0 (no range query) load d:read=0.95,insert=0.05,scan=0 (no range query) load e:scan=0.95,insert=0.05, range query, range 100 load f:read=0.5,readmodifywriteproportion=0.5 (read out record, then modify record, Then write back the record) test results for each indicator:

• IOPS Testing

Database type
Test Type	Mysql	Mongodb	Tokumx	Couchbase
Insert	13196.1154	77248.81715	36650.51286	92041.69321
A	17414.53382	120763.7097	56441.69009	171759.3308
B	68047.57887	99998.66668	98474.30477	159533.7361
C	65496.17721	96096.55782	96174.49901	97495.02775
F	16601.97431	83504.51759	58816.14856	112853.6551
D	66899.9244	100008.6674	97753.62177	101810.8761
E	2345.999594	2114.057191	2116.884482

• Insertion delay

	Database type
Test Type	Mysql	Mongodb	Tokumx	Couchbase
Insert	9683.304403	1658.783526	3490.779132	1337.107262
D	2238.196522	687.4064251	751.8563847	1246.501145
E	19863.49348	13568.54149	12914.24357

• Updsate delay

Database type
Test Type	Mysql	Mongodb	Tokumx	Couchbase
A	6840.743134	649.0976047	1670.448293	685.4414804
B	2260.907568	774.5944686	721.1753008	773.1773418
F	6583.488954	572.8773741	1198.842941	722.6498385

Read-Modify-write delay

	Database type
Test Type	Mysql	Mongodb	Tokumx	Couchbase
D	7133.568479	1050.091246	1677.069403	1442.553442

• Range Scan delay

Database type
Test Type	Mysql	Mongodb	Tokumx
E	24474.38263	31060.48241	31075.847

• Storage size (GB)

Database type
	Mysql	Mongodb	Tokumx	Couchbase
10 million	19	4.2	8.5	11.2

Conclusion: 1. Insert ops for all NoSQL databases read stronger than MySQL, where Couchbase is the strongest, then MongoDB and Tokumx 2. All NoSQL database insertion delay reads are smaller than MySQL, where Couchbase is the strongest and then MongoDB and Tokumx 3, read the most delay is couchbase, not as MySQL 4. Update delay, NoSQL is better than MySQL 5. Range Scan, MySQL is stronger than all NoSQL, due to couchbase test problem not getting data (subsequent additions) 6. Data capacity NoSQL database is smaller than MySQL, MySQL, couchbase, tokumx, MongoDB in sequence

YCSB Testing Mysql,mongodb,tokumx,couchbase Performance