I really do love your independence @Curbal. Even though you are an MVP, you voice your opinions freely, whether they be supportive, critical or cautious. Keep up the great work!
Best fabric video so far, it cuts through the MS marketing hype and clearly states some important risks. I work for a small charitable organisation with 150 power bi users. For the last few weeks it has felt as if we will have to move to fabric but your video puts my mind at rest . Just good old SQL and power bi for us 😅
Ruth is throwing cold water on my dreams. She raised my hopes with her one word, "Convenient", but then she added another word, "Danger". Let's just make a new word, "Convenient-danger". A word that strikes terror in the heart of my IT manager.
Thank you for the overview and insights! I was excited when fabric was announced, and started experimenting immediately. Within minutes, I realized how expensive it could be, having worked with the combined elements separately in the past. It does open up the potential role of data engineer to developers who haven't had the experience or exposure to these tools previously though. This may be part of the marketing plan ;)
My understanding is that “pay as you go” here means you have option to pause the capacity any time and save cost, and the max cost is when keep capacity running 24/7, similar cost to P capacity at reserved pricing. For SMB the lower F tiers will work better. The extra cost with OneLake storage can still be an issue but generally lake storage is cheap.
My understanding too. You have to pause it, which is not an easy thing to do when you spin off all those products. I have used Google pay as you go products and they can get really expensive if you are not careful.
Fabric sounds promising. While I do wish Microsoft would slow down with the rebranding, I find the current data toolset so difficult to configure and administer that id rather not use them. Hopefully with one unified platform, this becomes a lot easier. Thanks for this video Ruth, always appreciate the time and effort you put in to bringing us up to speed :)
The cost will be high with "pay as you go" pricing model, and it is the option for big company. But actually, I doubt that even big company will use Fabric when they already have their own data warehouse in some where else. As a BI person, I don't see the benefit that I will get from Fabric because I do not use notebook, do not take care warehouse, so why I have to pay for this, just integrate Power BI with Power Automate and Power App would be enough.
what enterprise databases are the target group? Are we talking about Oracle/SAP data warehouse reporting? Should this replace Oracle Analytics Cloud (OAC) for example? Currently we use OAC extracts, upload the files to SharePoint in order to report via PBI.
I am very confused what pay as you go means. If I use more compute then my F sku allows. What will happen. Will I pay more? Or will it become slow. The second is accetable, the first will mean a lot of management and control is necessary.
I havent tested the paid features yet, so I am not sure. On Google pay as you go means the more you pay literally for usage. I would assume here is the same but cant tell for sure.
Hello Ruth, My congratulations on your video and the simple and transparent way you put your opinion. I'm glad that being MVP didn't cloud your analysis and independence. 👏👏👏👏👏 As you mention, the goal of facilitating all data integration, necessarily has other motivations, as a private company this motivations are necessarily economic and domain. The price is effectively a black box, for now, and the surprises can be very unpleasant. The adoption of Microsoft Fabric by large companies will certainly lead to a large reduction in IT in these companies, and on the other hand, provide to these giant companies an even greater competitive advantage over smaller ones.
If Microsoft executes on their SaaS vision for Fabric, I think it is certainly a step forward from PaaS solutions. Platform services require a sgnificant level of admin overhead, planning, and tracking, and the interfaces are clunky and difficult to learn (as you mentioned). The question isn't what the absolute cost of Fabric is, but what the total cost of ownership is compared to other solutions that can deliver a simlar level of performance, scalability, and end-user efficiency. According to the OneLake videos that Microsoft has produced, my sense is that OneLake will address some of the metadata, discoverability, and lineage concerns that you have.
Hi Ruth, In the video you said something about connecting PBI to AWS. Do you know a way to get/connect to data stored in AWS S3 to PBI? Thank you! P.S. - Thank you for cautioning on how someone can quickly ramp up the Fabric bill!!!
For large companies adapting new tools take years time. So, no need to panic about Microsoft Fabric at the moment. But Data Analyst and Data Engineer roles may merge in couple of years I guess. Also AI becoming good at Data Analysis is a concern for core Data Analysts, becuase anyone can do Data Analysis with simple prompts. This landscape will shift by 2025 I guess.
Genuine question, not sarcasm but if you have your cloud infrastructure in multiple cloud platforms ar'nt you doubling your attack surface and adding complexity?
It is harder to attack two cloud services that just one? If you use only one cloud service you are completely exposed to their conditions. Minimum do a cloud and onprem in case you get in trouble with the cloud ! And different suppliers if you can.
On onelake, well we don’t just want one. We want different lakes for sensitive data segregation, security and archiving. I haven’t seen how that gets accommodated in fabric, but if it means onelake for analytics plus various other external lakes, well what’s the point.
The solution is what Microsoft are calling domains within OneLake. You can see them described in the Microsoft Developers video titled "Eliminate data silos with OneLake the OneDrive for Data."
I'm still have problems with new ribbon design in PBI Service. It looks a way more complicated even for non-rookies. And that was for Fabric unification purposes...
Great video Ruth. I thought the choice for a F SKU would mean some capping. But if not that would change it. I want to know beforehand what my cost will be. A Fabric PPU would be great :-)
@@CurbalEN the pay as you go is capped, as long as you turn off auto scaling. It allows you to reduce costs by turning it off and paying only for the time it is on. Also apparently you can use fabric with power bi premium capacity with no change in cost, it is the same as F64 SKU, except you can’t turn it off. This is likely what the reserved pricing will look like but with lower tier options that don’t get power bi premium features.
I mean the biggest for me is pricing model, which allows self service use the tools like spark notebook, Data Lake Gen 2, Synapse and Data Factory which allows to solve scenarios where Dataflows wheren't enough. Edit: Haven't watched till end of video. I didn't know it was Pay as you Go, I swear was thinking it was like Premium Capacity in PBI
I think the pricing model is terrible. Pay as you go is impossible to plan or budget for, specially when it comes to data. You never know what you will consume at any given point in time.
The outcome of companies adopting fabric is that PBI dev roles will get swallowed up by the data engineers/scientists. Visualisation is by far the easiest part of the whole BI function
Seriously, some whistleblower says the US Government is in possession of alien spacecraft. I would like to get my hands on whatever they use for data analysis. I hope all the technology is open source so we won't have to pay these Microsoft Fabric prices.
Very generous explanation, really. I expected to be surprised but not like this, this is really unmanaged, kinda scares me, after watching your entire summary. Tbh, this is no substitute for proper integrations, having a proper data lake, and then build data lakehouse implementation, with purpose built warehouses. I DO NOT want to intake excel sources from my boneheaded sales users, nor do I want deal with files being directly placed from the desktop to the lake (operated by the same boneheaded users) when there are perfectly functional MS solutions to build integrations such as SSIS or Azure Functions (or their counterparts in Google Functions or AWS Lambda) or external platforms such as Fivetran, Celigo or Boomi, with external, better, datalakes (Snowflake, BigQuery) This is just lazy. Convenience is how you end up with unscalable, ungoverned models which do NOT contribute to forming proper data models for.. anything, really.
The way I see, Microsoft just combined a bunch of services into one place. It's convenient not revolutionary like some people are trying to sell. And you still gonna have to pay by the hour if you wanna use it.
There's some rebranding and refreshing (updating). Having it all under 1 roof is a big deal. You could do a lot of this now without it but also you'd spend a lot more time to get all of it working as close to as smooth.
@@CurbalEN ❤️ Good review. Thanks. Microsoft has put too many marketing lights, sounds and whistles that when you step back to watch what they are offering it makes me feel like always: this is a new toy that the company won't afford.
Is Microsoft fabric really bringing value to a single user.. or is this all just about making money with Enterprise subscription.. typical Microsoft behavior... they are just going too fast.. building fear about AI😢 what do you think Ruth..
They have simplified the way you do data engineering for enterprises and create a new licensing model for that. When it comes to AI, they are going too fast with it without properly assessing the consequences. :(
Easier?? Honestly I hate MS products. 100000000 products without any real explanation. Almost all are duplicated etc, A huge confusion. No real manuals. Absolutely every MS product is a horror experience for me. Don't even talk about DAX...nightmare without any real MS manual.
I really do love your independence @Curbal. Even though you are an MVP, you voice your opinions freely, whether they be supportive, critical or cautious. Keep up the great work!
Thanks! ☺️
Great video! Unlike others that are promoting Fabric, it's great to hear different opinions in terms of what fabric will do for Power BI Developers.
Thanks :)
Best fabric video so far, it cuts through the MS marketing hype and clearly states some important risks. I work for a small charitable organisation with 150 power bi users. For the last few weeks it has felt as if we will have to move to fabric but your video puts my mind at rest . Just good old SQL and power bi for us 😅
Sounds just right. Glad the video helped!
Ruth is throwing cold water on my dreams. She raised my hopes with her one word, "Convenient", but then she added another word, "Danger". Let's just make a new word, "Convenient-danger". A word that strikes terror in the heart of my IT manager.
Covenient danger 😂😂 It applies to so many things!!
Really appreciate the candid and transparent thinking you provide here. The So-What implications really jump out.
Thanks Vijay :)
Thank you for expressing an opinion about this product in an honest and unbiased way
Thanks Andrea, I always try to be fair in my reviews :)
Very informative and honest video, which is rare these days.
Thank you!
Thanks 2 u!
Longtime viewer, first time commenting.
Your content is incredible, thank you.
Thanks 😃
Great video... but now I'm intrigued to find out how horrible the data lake setup is.. just for comparison. Mind to inform which video it is ? Thanks.
Thank you for the overview and insights! I was excited when fabric was announced, and started experimenting immediately. Within minutes, I realized how expensive it could be, having worked with the combined elements separately in the past. It does open up the potential role of data engineer to developers who haven't had the experience or exposure to these tools previously though. This may be part of the marketing plan ;)
3:30 what's the name/link to your other video?
Here:
m.ruclips.net/video/g73TwnjjTZw/видео.html
@@CurbalEN Thank you so much. Btw, we appreciate a lot your informed and honest take on new releases!!
thank you for this overview. I l ove how you always keep it real and lay it out as it is.
Thanks Nate :)
17:40 IF the company is NOT already paying for any Premium license/Capacity it won't do it for Fabric.
My understanding is that “pay as you go” here means you have option to pause the capacity any time and save cost, and the max cost is when keep capacity running 24/7, similar cost to P capacity at reserved pricing. For SMB the lower F tiers will work better. The extra cost with OneLake storage can still be an issue but generally lake storage is cheap.
My understanding too. You have to pause it, which is not an easy thing to do when you spin off all those products. I have used Google pay as you go products and they can get really expensive if you are not careful.
Fabric sounds promising. While I do wish Microsoft would slow down with the rebranding, I find the current data toolset so difficult to configure and administer that id rather not use them. Hopefully with one unified platform, this becomes a lot easier. Thanks for this video Ruth, always appreciate the time and effort you put in to bringing us up to speed :)
Fabric has definitely done a great job to make those data engineering products easier to use :)
Thank you, this was the best Fabric video that answer all my questions.
😊😊
Can a on-prem SQL server be set up as a shortcut by using a remote gateway? This would ensure that our data at rest would still be on-prem.
Good question… my guess would be only cloud services, but I could be wrong. Ask in the community to be sure!
The cost will be high with "pay as you go" pricing model, and it is the option for big company. But actually, I doubt that even big company will use Fabric when they already have their own data warehouse in some where else.
As a BI person, I don't see the benefit that I will get from Fabric because I do not use notebook, do not take care warehouse, so why I have to pay for this, just integrate Power BI with Power Automate and Power App would be enough.
If your organisation sets up oneLake you probably will use that, for the rest, we are in agreement :)
Thanks Ruth for the great video.
Could you please put the link for the Microsoft video you reffering to, on your video?
I checked and it seems to be there:
Pricing viceo here:
ruclips.net/video/6AAeV3bSMso/видео.html&ab_channel=GuyinaCube
@@CurbalEN
Thanka a lot!
But I meant the other one. The one that the lady from microsoft explein about the Fabric platform.
Oh, here it is:
m.ruclips.net/video/oxMUJT3p9f4/видео.html&vl=en
Will you be covering fabric in more detail with use cases in future ?
Not likely, but if something really cool pops up I will :)
I like what you said in the end: "Do not put all your eggs in one basket"
what enterprise databases are the target group? Are we talking about Oracle/SAP data warehouse reporting? Should this replace Oracle Analytics Cloud (OAC) for example? Currently we use OAC extracts, upload the files to SharePoint in order to report via PBI.
At the moment they have integration with Amazon S3. I am sure they will
Extend that to other clouds.
So different points that touch the data (e.g. onedrive). Does this add an issue around data governance?
You will definitely need to think through and plan ahead how to use and deploy Fabric.
I am very confused what pay as you go means.
If I use more compute then my F sku allows. What will happen. Will I pay more? Or will it become slow. The second is accetable, the first will mean a lot of management and control is necessary.
I havent tested the paid features yet, so I am not sure. On Google pay as you go means the more you pay literally for usage. I would assume here is the same but cant tell for sure.
Hello Ruth,
My congratulations on your video and the simple and transparent way you put your opinion.
I'm glad that being MVP didn't cloud your analysis and independence. 👏👏👏👏👏
As you mention, the goal of facilitating all data integration, necessarily has other motivations, as a private company this motivations are necessarily economic and domain.
The price is effectively a black box, for now, and the surprises can be very unpleasant.
The adoption of Microsoft Fabric by large companies will certainly lead to a large reduction in IT in these companies, and on the other hand, provide to these giant companies an even greater competitive advantage over smaller ones.
Lets see how the adoption goes, too early to tell !
As a new Data Architect who was last a Power BI Developer, I wish Fabric came out 3 years ago. The future is focused on Power BI
Provisioning and connecting takes hours and months waiting for other things to be provisioned
If Microsoft executes on their SaaS vision for Fabric, I think it is certainly a step forward from PaaS solutions. Platform services require a sgnificant level of admin overhead, planning, and tracking, and the interfaces are clunky and difficult to learn (as you mentioned). The question isn't what the absolute cost of Fabric is, but what the total cost of ownership is compared to other solutions that can deliver a simlar level of performance, scalability, and end-user efficiency. According to the OneLake videos that Microsoft has produced, my sense is that OneLake will address some of the metadata, discoverability, and lineage concerns that you have.
Fingers crossed they will!!
Is it like an ERP?
No, more like data engineering tools brought together
Hi Ruth,
In the video you said something about connecting PBI to AWS.
Do you know a way to get/connect to data stored in AWS S3 to PBI?
Thank you!
P.S. - Thank you for cautioning on how someone can quickly ramp up the Fabric bill!!!
They have a REST API
docs.aws.amazon.com/AmazonS3/latest/API/Welcome.html
And 😂😂
For large companies adapting new tools take years time. So, no need to panic about Microsoft Fabric at the moment. But Data Analyst and Data Engineer roles may merge in couple of years I guess. Also AI becoming good at Data Analysis is a concern for core Data Analysts, becuase anyone can do Data Analysis with simple prompts. This landscape will shift by 2025 I guess.
The skills required for both roles are fundamentally different, so I dont see a merge any time soon, but AI will definitely disrupt both roles.
Genuine question, not sarcasm but if you have your cloud infrastructure in multiple cloud platforms ar'nt you doubling your attack surface and adding complexity?
It is harder to attack two cloud services that just one?
If you use only one cloud service you are completely exposed to their conditions. Minimum do a cloud and onprem in case you get in trouble with the cloud ! And different suppliers if you can.
On onelake, well we don’t just want one. We want different lakes for sensitive data segregation, security and archiving. I haven’t seen how that gets accommodated in fabric, but if it means onelake for analytics plus various other external lakes, well what’s the point.
Yeah, it maybe is a onelake for data engineering projects. Not sure either…
The solution is what Microsoft are calling domains within OneLake. You can see them described in the Microsoft Developers video titled "Eliminate data silos with OneLake the OneDrive for Data."
Thanks, will watch that :)
I'm still have problems with new ribbon design in PBI Service. It looks a way more complicated even for non-rookies. And that was for Fabric unification purposes...
It is more complicated.
Distribute your reports through apps only and keep your business users away from Power BI service. Best practice imo.
Great video Ruth. I thought the choice for a F SKU would mean some capping. But if not that would change it. I want to know beforehand what my cost will be. A Fabric PPU would be great :-)
I can tell you 100% but pay as you go usually is pay by usage. Happy to retract if that is not the case :)
@@CurbalEN the pay as you go is capped, as long as you turn off auto scaling. It allows you to reduce costs by turning it off and paying only for the time it is on. Also apparently you can use fabric with power bi premium capacity with no change in cost, it is the same as F64 SKU, except you can’t turn it off. This is likely what the reserved pricing will look like but with lower tier options that don’t get power bi premium features.
@dodgysuspect7103 Thanks!!
Det var mer än ett ord men tack för en jättebra genomgång. Det är DWH som är dyr och den borde man kunna ändra till vanlig Azure SQL Server.
Ordet är ett, men förklaring av varför just den kräver några till :)
@@CurbalEN 💛 Väl spenderade 18-19 minuter.
Tack 💛
I mean the biggest for me is pricing model, which allows self service use the tools like spark notebook, Data Lake Gen 2, Synapse and Data Factory which allows to solve scenarios where Dataflows wheren't enough.
Edit: Haven't watched till end of video. I didn't know it was Pay as you Go, I swear was thinking it was like Premium Capacity in PBI
I think the pricing model is terrible. Pay as you go is impossible to plan or budget for, specially when it comes to data. You never know what you will consume at any given point in time.
@@CurbalEN I not sure I follow Adam said in that moment that PBI Premium got you covered ruclips.net/video/6AAeV3bSMso/видео.html
Nice!
The outcome of companies adopting fabric is that PBI dev roles will get swallowed up by the data engineers/scientists. Visualisation is by far the easiest part of the whole BI function
Or power bi engineers will swallow data engineers? It is the power bi platform you guys are stepping in ;)
Both ways are good news :)@@CurbalEN
Seriously, some whistleblower says the US Government is in possession of alien spacecraft. I would like to get my hands on whatever they use for data analysis. I hope all the technology is open source so we won't have to pay these Microsoft Fabric prices.
The 80s all over again!! 😂
theres a little error in the title "farbric"
Or is it 🧐 lol
Thanks! I fixed it :)
😂😂
One word... 'Expensive'
It could be if you are no careful ….
Very generous explanation, really.
I expected to be surprised but not like this, this is really unmanaged, kinda scares me, after watching your entire summary.
Tbh, this is no substitute for proper integrations, having a proper data lake, and then build data lakehouse implementation, with purpose built warehouses.
I DO NOT want to intake excel sources from my boneheaded sales users, nor do I want deal with files being directly placed from the desktop to the lake (operated by the same boneheaded users) when there are perfectly functional MS solutions to build integrations such as SSIS or Azure Functions (or their counterparts in Google Functions or AWS Lambda) or external platforms such as Fivetran, Celigo or Boomi, with external, better, datalakes (Snowflake, BigQuery)
This is just lazy. Convenience is how you end up with unscalable, ungoverned models which do NOT contribute to forming proper data models for.. anything, really.
Good point, data environment must be carefully planned. Letting anyone do everything will lead to chaos!
Hype. That's the word.
me and my data engineer pal will build our own lakehouse with blackjack and hookers (read: APIs, Talend & Pentaho) 🤣🤣🤣
The way I see, Microsoft just combined a bunch of services into one place. It's convenient not revolutionary like some people are trying to sell. And you still gonna have to pay by the hour if you wanna use it.
It is a big step for them, but not for the data world imo.
There's some rebranding and refreshing (updating). Having it all under 1 roof is a big deal. You could do a lot of this now without it but also you'd spend a lot more time to get all of it working as close to as smooth.
Convenience? Is the word convenience?
It is :)
Nice
It seems like a money vacuum - that giant sucking sound is your dollars going to Redmond.
If you are not careful 😂😂
An 18 minute video to explain it in ONE WORD 😂
It is the opposite, I explained one word in an 18 min video. 😉
13:00 agree with you.
🔥
Before start. In one word? But it needs a 19 minutes video to explain it
Good point, should have said summarise in one word! 😂
@@CurbalEN ❤️ Good review. Thanks. Microsoft has put too many marketing lights, sounds and whistles that when you step back to watch what they are offering it makes me feel like always: this is a new toy that the company won't afford.
Is Microsoft fabric really bringing value to a single user.. or is this all just about making money with Enterprise subscription.. typical Microsoft behavior... they are just going too fast.. building fear about AI😢 what do you think Ruth..
They have simplified the way you do data engineering for enterprises and create a new licensing model for that.
When it comes to AI, they are going too fast with it without properly assessing the consequences. :(
You need to be extremely careful before you like people like me play around with this.
We are twin souls, i would be extremely dangerous too 😂
Easier?? Honestly I hate MS products. 100000000 products without any real explanation. Almost all are duplicated etc, A huge confusion. No real manuals. Absolutely every MS product is a horror experience for me. Don't even talk about DAX...nightmare without any real MS manual.
🥺