Announced in early 2023, Fabric is the name being given to a largely existing suite of applications within Azure (and O365). Bringing all data services under a unified heading. There are simplifications and deduplication within the structure as (currently) separate workloads can be unified for more business benefit.
So, where’s the rub? Let me get the elephant in the room out of the way immediately: for me and for all the clients I have discussed Fabric with, it is cost! The only indication that I can get so far is that you need to have a “Capacity” subscription, so Power BI Premium capacity is an example. To be candid, while this is a significant cost, it would not be as bad if the support requirements were clarified. In other words, the cost of capacity can be offset by not needing as many support resources if the business can be persuaded or can take on their share of the development activities. The challenge here again is that the platform is still too fluid for it to be clear that this is something that can reasonably be assessed.
Do not be put off though by what I am saying the promise that Fabric is so far bringing makes is something that I would definitely recommend, and I have no doubt that a lower-cost option will become available in the coming months. While there is a risk of this being considered as me revealing the secrets of the Magic Circle, the cloud has so far proven difficult for Microsoft to make available to consultancies. Very few consultancies will pay for Power BI Premium Capacity, and I am sure that was a factor in Premium Per User being launched within Power BI, something that has now become the defacto deployment standard for all of our engagements despite us trying to steer customers towards Pro ONLY. The updated Data Flow capabilities often make it more than worth the extra cost. The benefit of this change is that the friction of a move to Premium Capacity is greatly lowered but you still need to have around 200 Power BI users to make it worth the switch OR 500+ users you wish to be able to consume simple report content. So, clarity of deployment shapes the roadmap for clients moving into Power BI, with many smaller to medium-sized organisations simply ruling Premium Capacity subscriptions either out of reach or not a fiscal priority. Drawing this together a significant proportion of Power BI deployments and the consultancies that support Power BI deployments will not be getting adequate experience once the Preview is completed (imminent at the time of writing). So the cost will have to come down!
Some pricing is starting to come out in terms of Pay-as-you-go, but the exact capability of any tier on the table just is not clear, especially when it is considered that this relates to processing rather than storage. In short, there are still more questions raised than answered.
So why should I bother to look at anything about Fabric now? Well honestly, Fabric is a Data Culture focused solution to your business, meaning that it gives the capability of setting up and unifying your datastores, so your line of business and forecasting models are aligned and use the same data. This has been the direction Microsoft have been steering us for many years from the first mention of the “Common Data Service”, finally technology has caught up to allow the concept to be more than just a set of (often unpopular) rules – I’ve lost count of the number of times I’ve been told a business’s version of a table is completely incompatible with the Common Data Service.
The improved Data Flows and Data Factory alone make an extremely compelling argument in favour of looking more deeply into Fabric. Access to all data via SQL, Python or the Power BI engine means existing solutions can be migrated easily while being able to report more rapidly from the production systems further lowering the Total cost of Ownership for the organisation. OneLake will make it so you only need to store the data once and then use compute resources (costs may necessitate that you process in batches and store multiple times – as we do now). Security is simplified the fewer times data is stored, so having the capability to only store data once and then access it across the raft of Fabric outputs is a must.
To conclude the difference between a platform that will be the heart of your businesses data or the Emperor’s new Clothes will be largely dependent on the pricing model and that being understandable by those outside of IT. Fabric after all is intended to bring the business to the fore in terms of solution delivery with the traditional IT organisation providing the guide rails to keep the business and their data safe.
