Pricing changes are coming in January. Contact our Sales team to secure the current price for your desired license.
Get Free Trial
Get Free Trial

Increase performance working with MS Analysis services data

posted on June 30th, 2024

Answered
Yosef asked on June 30, 2024

We are connecting to MS Analysis Services cubes with a fair amount of data through our own XMLA handler. In some report configurations it's becoming slow to load, both the members and the actual data, and we can achieve a similar query connected to the same endpoint with Excel comparably faster.

What do you suggest we do to improve this?

Would there be a performance difference between using a XMLA proxy vs Flexmonster Accelerator? What if both are located on the same server as the cube?

In terms of the distribution of the requests we can see there are several requests made when loading the members and then a single big one for the data - between them is where the time gets consumed. One concern here is extra time taken for the individual member requests with latency for each request between the client and the server. Is there anything we can do to influence this from our end? 

2 answers

Public
Maksym Diachenko Maksym Diachenko ⋅ Flexmonster ⋅ July 1, 2024

Hello, Yosef!

Thank you for reaching out to us.

Flexmonster Accelerator helps to increase the data loading speed when working with cubes in Flexmonster. The performance improvement with the Accelerator can be attributed to its following characteristics:

  • Custom communication protocol: Instead of loading XML responses, Flexmonster Accelerator sends the data to Flexmonster in JSON format. This improves the loading time via the network because fewer characters are used to transfer the data. Additionally, JSON responses are parsed faster than XMLA due to a simpler structure. Please check the network requests within this JSFiddle to see the example of JSON payloads when connected to the Accelerator.
  • Caching: Flexmonster Accelerator provides an opportunity to cache responses, which helps to improve performance by loading in-memory data instead of querying the cube.

Running the Accelerator on the same server as the cube is important since this ensures the absence of network latency between different servers.

You can check our documentation regarding different installation approaches for Flexmonster Accelerator:

We are looking forward to hearing your feedback.

Best Regards,
Maksym

Public
Maksym Diachenko Maksym Diachenko ⋅ Flexmonster ⋅ July 23, 2024

Hello, Yosef!

Hope you are doing well.
We are wondering if you have tried using Flexmonster Accelerator to connect your cube to Flexmonster. If so, did it help improve the loading speed?
Looking forward to hearing from you.

Best Regards,
Maksym

Please login or Register to Submit Answer