Forum Discussion
Which causes greater load stress?
Our sysadmin wants to use one giant power bi model (fed by dataflows and in import mode) to create reports from. Basically, his idea is build a giant model and have people build reports from this dataset in the service. Or via live connection if in the desktop.
The other option is that we would break datasets down into core subject matter data sets. These would be composite models (combination of dataflows and direct query datasets). Users in the cloud would build reports of whichever particular dataset fit their business need.
His concern with the 2nd option is that he thinks it will put a lot of stress on our ERP system and Salesforce (the two main systems we use and that the dataflows are based on). We are not a huge company, and have less than 300,000 records.
Just curious which of the above scenarios is the least stressful for our backend systems? In terms of performance, visuals, page loading, etc
Hi Anonymous,
If you consider reducing the load on the back-end system, using import mode is definitely a good idea, but considering the need to import data into Power BI, importing data and refreshing it will consume a lot of time if the data source is updated frequently.
Please refer Optimization guide for Power BI.
https://docs.microsoft.com/en-us/power-bi/guidance/power-bi-optimization
If you still have some question, please don't hesitate to let me known.
Best Regards,
Link
Is that the answer you're looking for? If this post helps, then please consider Accept it as the solution. Really appreciate!
4 Replies
Anonymous , one of the option can be explored is dataflows and then direct query on dataflows. That can keep your model light. (This is a preview, can have some performance issues)
Having one datafile having everything import is not a good idea. Have dataflows, Datasets and try to use them
amitchandak Can you elaborate a little on why it's not a good idea? Just looking for some technical reasons why the one giant data model is not a good option.
My preference is to use dataflows (in import mode, not DQ). DQ has too many limitations in terms of the data transformations etc.
Anonymous , If you want complete import the having a single model as dataset can save space. As long as you can push data(full) to loaded on power bi service(deployement pipeline)
If you can use direct query then data is loaded at one place , you can work wih smaller model as per need.
Hi Anonymous,
If you consider reducing the load on the back-end system, using import mode is definitely a good idea, but considering the need to import data into Power BI, importing data and refreshing it will consume a lot of time if the data source is updated frequently.
Please refer Optimization guide for Power BI.
https://docs.microsoft.com/en-us/power-bi/guidance/power-bi-optimization
If you still have some question, please don't hesitate to let me known.
Best Regards,
Link
Is that the answer you're looking for? If this post helps, then please consider Accept it as the solution. Really appreciate!
