Quantcast
Channel: Randy Riness @ SPSCC aggregator
Viewing all articles
Browse latest Browse all 3015

MSDN Blogs: Master Data Services 2016 Performance Enhancements

$
0
0

MDS 2016 includes many improvements to performance and capacity. This document shares some early performance test results that compare MDS 2014 with pre-release MDS 2016.

 

The tests were run on an Azure GS3 Virtual Machine (8 cores, 112 GB memory, 1 TB premium storage). The test data consists of a model with 7 entities and 2 versions. The largest entity contains 7 million members and 18 attributes, 6 of which are domain-based.

 

 

MDS 2014 SP1

MDS 2016 (pre-release)

Operation

Time

Time

Versus MDS 2014

Copy version

1:01:25.0

0:29:17.0

210%

 

 

 

 

Metadata operations

 

 

 

Get model details

0:00:02.9

0:00:00.6

483%

Create entity with 7 attributes

0:00:03.3

0:00:01.2

275%

Create entity with 200 attributes

0:01:02.2

0:00:02.1

2,962%

 

 

 

 

Master data operations (using Excel add-in)

 

 

 

Load 1,000 members (using filter)

0:00:01.1

0:00:00.5

220%

Load 50,000 members (using filter)

0:00:11.7

0:00:06.5

180%

Create 1,000 members

0:00:05.2

0:00:03.0

173%

Create 50,000 members

0:03:49.9

0:02:07.1

180%

Update 1,000 members

0:00:05.9

0:00:03.5

169%

Update 50,000 members

0:03:45.1

0:02:21.0

160%

Delete 1,000 members

0:00:02.7

0:00:00.6

450%

 

 

 

 

Model deployment operations

 

 

 

Create Package

N/A (30+ hours)

0:35:19.0

 

Deploy Package (Update)

N/A

5:12:40.0

 

Deploy Package (Clone)

N/A

4:19:02.0

 

 

MDS 2014 could not complete the model deployment tests because the data set was too large for it to handle. The Create Package operation did not complete after running for more than 30 hours.

 

The release-to-market (RTM) version of MDS 2016 includes even more performance improvements. Stay tuned for details!


Viewing all articles
Browse latest Browse all 3015

Trending Articles