Power BI Deployment

Power BI Deployment Question

Given the below company criteria, please can users make recommendations on the optimal Power BI reporting structure. With the introduction of Power BI Premium User licences, I wanted to see if anyone has a better idea for deployment / setup?

Current Setup

• 120 Power BI Pro users (all using internal domain address accounts). Split between internal users (70%) and external users (30%)
• Internal Users are split into 5 BU’s which do not have shared access to workspaces
• External Users are split into 4 groups who also cant see each others data
• IT is responsible for report creation driven by business requests
• Each user internal / external has a Power BI Pro licence
• The information available to external users will be different from the information used by internal
• Most users are “viewers” of reports (they change a few filters and that’s it) with small % of users requiring exports of data which is then manipulated in excel (multiple data silos I know). No users are building their own models / creating measures etc.
• The data is sourced from a non-azure data warehouse and consists of multiple datasets being repeated and inefficient
• The market leading technology is adopted rather than focussing entirely on Microsoft universe. Hence Sharepoint is not utilised despite the Office 365 availability (I mention this as Sharepoint is one change I have considered for PBI report roll out)

Proposed Setup

• Dataflow approach instead of multiple datasets
• 5-10 internal users retaining Pro/Upgraded to premium user licenses
• External users provide their own licenses and data shared externally (via app?)
• Majority of internal users view data through some embedded option. Can not be published to web as needs to be secure (advice on embedded options / costing would be appreciated)

Thanks in advance and I look forward to your feedback

Hi All,

Anything wrong with my posting as no responses. Any advise on embedding / power bi premium user options would be massively appreciated

Thanks

@Hitman,

No, the question was really well framed. Unfortunately, I’ve only worked within the “one pro license per user” environment, so I don’t have experience on this one to provide you a good response. I’m looping in @datazoe, @Melissa and @haroonali1000 , who I think are the most knowledgeable among the expert team on deployment options.

  • Brian

Thanks Brian. As always your assistance massively appreciated. Hope you are keeping well… Andrew

I look forward to hearing from @datazoe, @melissa, @ haroonali1000.

It is interesting that perhaps the most important part (deployment) is not an area that many sources provide detail on. I know the report server and embedding options are a bit of a grey area to me…

Perhaps another potential video course option for EDNA (your welcome ;)) but for now any feedback from users on the highlighted scenario would be massively appreciated

Hi Andrew (@Hitman)!

With 120 pro you are paying $1200 monthly, and if you did go PPU it would be double per month, so $2400. With PPU you have to give both publishers and viewers licenses to access any workspace put in the PPU capacity. Even if it’s just IT publishing reports, the workspace is what gets flagged as PPU (by any PPU user) and then only PPU users can publish or view content.

That said, $2400 is still much cheaper than $5000/mo + pro licenses for the publishers (which from your scenario is less than your 120) for P sku premium (per org) capacity. Again it’s the workspace would also need to put in the capacity (by Power BI Admin) for it to be viewable to anyone in the org without a pro license. So you may need to still keep existing users with pro licenses if they are publishing reports.

As for why you should double the cost:

  1. Deployment pipelines in Power BI! I love these and I suspect IT may also. It creates a DEV, TEST, and PROD workspace and it works really well if you have a dev and prod data source going on. You have many disparate sources, so another benefit is as you consolidate them (if that is a plan) you can do so in the dev workspace and then be able to compare dev to prod side by side after published. You can do it other ways, but the pipelines allow you to simply deploy to the next stage and it lets you know when things are out of sync and other goodness. Deployment pipelines, the Power BI Application lifecycle management (ALM) tool, process - Power BI | Microsoft Docs
  2. Goals, the new thing, may be helpful. Introducing Goals in Power BI | Microsoft Power BI Blog | Microsoft Power BI
  3. Paginated reports, as always the fav premium feature.
  4. Publishers can utilize ALM toolkit to update in place data sets versus republishing the whole report (IT may like this one). Home Page - ALM Toolkit (alm-toolkit.com)
  5. the data model is surfaced as XMLA endpoint in premium, so you can directly query it in DAX via DAX Studio or SSMS and it can be used in other tools such as Spotfire. You can also directly query in Excel instead of using Analyze in Excel, and that’s so much faster in many cases! Dataset connectivity and management with the XMLA endpoint in Power BI - Power BI | Microsoft Docs
  6. Cognitive Services is now at your disposal in Power Query – Sentiment Analysis of text fields, Key Phrase extraction, and Language Detection. Tutorial: Use Cognitive Services in Power BI (Preview) - Power BI | Microsoft Docs
  7. I’m sure there are other benefits but these are the ones I use. What is Microsoft Power BI Premium? - Power BI | Microsoft Docs

There are also A premium skus for premium per org in the Azure marketplace too, and they start at around $1200ish per month (I think, I would double check that one!). They are very small, so would only work on very small models.

Another thing to note is P skus are 25GB but that’s the memory for refreshing and at rest, that’s to say, you can’t put 25GB of data models on it! Max I’d say is around 5GB. When datasets refresh they tend to spike in memory usage, then once done go down. The 25GB has to be enough for the spikes too.

Hope that helps! I don’t know much about Data Flows to be honest – I like to create big data models in PBIX files then publish them up as the “central” or “golden” dataset that feeds everything else (you can promote and certify to increase visibility). You can segment the data inside that one model via Row Level Security for different business units or specific users too.

5 Likes

@datazoe ,

Fantastic - thanks! That one is definitely getting bookmarked.

@Hitman - reading your earlier response jogged my memory on some other resources that will probably be very helpful as well. @sam.mckay has put together a number of excellent licensing and deployment guides that you can download in the area of the portal highlighted below:

In conjunction with Zoe’s info above, should be quite useful.

FYI - re: Dataflows, I think Experts @Heather and @Melissa are both knowledgeable in this area.

  • Brian

Hi @datazoe

Some food for thought there so thank you very much. I have to admit i didn’t realise pipelines and goals was specifically for premium as i recently upgraded and these i think were released about the same time. I assumed they were standard.

Here are my thoughts and it would be interesting to know if you agree:

  1. IT developers should always be Prem users as for the small incremental cost it allows a) pipeline functionality b) ALM Toolkit c) XMLA endpoint functionality

  2. Unless you roll out across out across the whole organisation (evenif you had a couple of power bi user licenses in IT) you cannot take advantage of the functionality paginated reporting and goals as it would require the end user to have the prem user licence- True?

  3. I thought that a good solution for the mentioned scenario would be to have a couple of licences for IT and to embed the reports into a sharing platform (i have heard of companies doing this but have little detail on it). This would remove the license costs and also meet the functionality of users in this scenario. I believe there would be report server costs but I know little about this and how it would compare to each user having a pro license.

I also know nothing of the cognitive services side so will enjoy reading this - Thank you

Thank you again for your response and will be going through the points in further detail

1 Like

Thanks Brian - I will go through these (i think maybe i have but am not 100%) but i dont know if they were updated for the prem user side or if it discusses about embedding. I think the problem with this area that is that there have been some pretty big changes to it recently

@Hitman ,

Unfortunately, correct - I just checked and these haven’t yet been updated for 2021. I’ll talk with Sam on Monday and see what the plans for these are. That’s both one of the best and one of the most difficult things about PBI - you’re always chasing a moving target…

  • Brian

Hi @Hitman, did the response provided by the Experts help you solve your query? If not, how far did you get and what kind of help you need further? If yes, kindly mark as solution the answer that solved your query. Thanks!

Hi @HItman, we’ve noticed that no response has been received from you since the 30th of May. We just want to check if you still need further help with this post? In case there won’t be any activity on it in the next few days, we’ll be tagging this post as Solved.

Hi @HItman, due to inactivity, a response on this post has been tagged as “Solution”. If you have a follow question or concern related to this topic, please remove the Solution tag first by clicking the three dots beside Reply and then untick the check box.

Kindly take time to answer the Enterprise DNA Forum User Experience Survey, we hope you’ll give your insights on how we can further improve the Support forum. Thanks!

@Hitman,

Glad you found the information useful! Yes, if you rolled out PPU workspaces to all org then whole org would need PPU to see it. You can alternatively have PPU workspace with pipelines for just working through model changes, etc, then when you are happy, publish it as a next step to a pro workspace. Then pro users can access it there. Paginated and goals though will have to stay in the PPU workspace.

For embedded, that’s actually using the A-skus in Azure, which is then premium per org, so any workspace in a premium capacity can simply be shared to all org (pro license or not can access). Embedded has added bonus of sharing reports to say a portal your cilents log into and then in the dev code that calls Power BI for the report you specify who it is and roles you set up can segment data (or not if all data is fair game!). You would need dev skills to pull this off! More info available here: New Contoso Sample app for Power BI Embedded - YouTube which also included more links including the Power BI Playground Power BI Playground.

Hope this helps!

2 Likes

Thanks @datazoe (apologies for delay this week).

1 Like