July 22, 2026
Microsoft Fabric - Simply Explained
Data has become one of every organization's most valuable assets—but for many businesses, it's scattered across databases, cloud platforms, business applications, and analytics tools. Microsoft Fabric was created to solve this fragmentation by bringing every stage of the data lifecycle into a single, unified platform. In this episode of Microsoft Knowledge Nuggets, we explain Microsoft Fabric in plain English, exploring how it combines data engineering, analytics, business intelligence, artificial intelligence, and governance into one modern Software-as-a-Service platform.
WHY MICROSOFT CREATED FABRIC
For decades, organizations managed separate tools for data storage, ETL pipelines, analytics, reporting, machine learning, and business intelligence. Every system required its own infrastructure, administration, security model, and data movement. As data was copied between platforms, businesses created multiple versions of the truth while spending more time moving data than analyzing it. Microsoft Fabric eliminates these silos by providing a single platform where data is stored once and used everywhere.
UNDERSTANDING ONELAKE: THE FOUNDATION OF MICROSOFT FABRIC
At the heart of Microsoft Fabric is OneLake, a single enterprise-wide data lake that acts as one centralized source of truth for the entire organization. Similar to how OneDrive stores personal files, OneLake stores business data in open Delta Parquet formats while allowing every Fabric workload to access the same information without duplication. Features such as shortcuts even allow organizations to reference data stored in AWS, Google Cloud, or on-premises environments without physically moving it, simplifying hybrid and multi-cloud architectures.
LAKEHOUSES, WAREHOUSES, AND DATA ENGINEERING
Microsoft Fabric supports multiple ways of working with data depending on your role. Data engineers can build scalable pipelines inside Lakehouses using Spark, Python, notebooks, and Delta tables. SQL professionals can work inside fully managed Warehouses that provide familiar T-SQL experiences while accessing the exact same data stored in OneLake. Since both workloads share a common storage layer, organizations avoid unnecessary copies while enabling collaboration between engineering, analytics, and business intelligence teams.
DATA FACTORY, REAL-TIME ANALYTICS, AND ARTIFICIAL INTELLIGENCE
Fabric includes Microsoft Data Factory with hundreds of built-in connectors, enabling organizations to ingest data from virtually any source. Dataflows Gen2 simplify transformation using Power Query, while Mirroring and OneLake Shortcuts reduce the need for complex ETL pipelines. Fabric also supports real-time analytics through Eventhouses and Kusto Query Language (KQL), allowing organizations to process streaming IoT, telemetry, and operational data in near real time. Built-in AI capabilities and Microsoft Copilot further accelerate report creation, SQL generation, pipeline development, and machine learning by allowing users to interact with their data using natural language.
POWER BI, GOVERNANCE, AND ENTERPRISE DATA MANAGEMENT
Power BI is deeply integrated into Microsoft Fabric through Direct Lake mode, enabling reports to query OneLake directly without importing or duplicating data. Semantic models, web-based development, Microsoft Purview governance, sensitivity labels, data loss prevention policies, and domain-based administration help organizations maintain strong security while empowering business users with trusted, governed analytics. Fabric's unified governance model ensures data remains protected throughout its entire lifecycle while simplifying compliance and enterprise data management.
WHY MICROSOFT FABRIC MATTERS
Microsoft Fabric isn't simply another analytics product—it represents Microsoft's vision for a unified data platform where storage, engineering, analytics, AI, governance, and business intelligence work together seamlessly. By eliminating data silos, reducing infrastructure complexity, and enabling organizations to store data once while analyzing it everywhere, Fabric helps businesses accelerate decision-making, improve collaboration, and unlock greater value from their enterprise data.
Become a supporter of this podcast: https://www.spreaker.com/podcast/m365-fm-modern-work-security-and-productivity-with-microsoft-365--6704921/support.
🚀 Want to be part of m365.fm?
Then stop just listening… and start showing up.
👉 Connect with me on LinkedIn and let’s make something happen:
- 🎙️ Be a podcast guest and share your story
- 🎧 Host your own episode (yes, seriously)
- 💡 Pitch topics the community actually wants to hear
- 🌍 Build your personal brand in the Microsoft 365 space
This isn’t just a podcast — it’s a platform for people who take action.
🔥 Most people wait. The best ones don’t.
👉 Connect with me on LinkedIn and send me a message:
"I want in"
Let’s build something awesome 👊
1
00:00:00,000 --> 00:00:01,520
Welcome to another episode.
2
00:00:01,520 --> 00:00:03,360
Today's topic is one you've probably heard
3
00:00:03,360 --> 00:00:05,000
in meetings or tech articles.
4
00:00:05,000 --> 00:00:07,480
Microsoft fabric, but if you're like most people,
5
00:00:07,480 --> 00:00:09,680
you're not entirely sure what it actually is.
6
00:00:09,680 --> 00:00:11,360
Is it just another Microsoft product,
7
00:00:11,360 --> 00:00:12,600
a rebrand of something old,
8
00:00:12,600 --> 00:00:14,320
or is it something completely different?
9
00:00:14,320 --> 00:00:15,800
Let me give you the short answer.
10
00:00:15,800 --> 00:00:17,760
Microsoft fabric is not just another product.
11
00:00:17,760 --> 00:00:19,040
It's a complete data platform,
12
00:00:19,040 --> 00:00:21,240
one place where you can do all your data work,
13
00:00:21,240 --> 00:00:24,400
storage, transformation, analytics, and reporting,
14
00:00:24,400 --> 00:00:25,440
all in one place.
15
00:00:25,440 --> 00:00:26,680
By the end of this episode,
16
00:00:26,680 --> 00:00:29,400
you'll understand what fabric is, why Microsoft built it,
17
00:00:29,400 --> 00:00:31,240
and how the main pieces fit together.
18
00:00:31,240 --> 00:00:33,440
Grab your coffee and let's dive in.
19
00:00:33,440 --> 00:00:36,320
The old way, why data was broken?
20
00:00:36,320 --> 00:00:38,360
To understand why fabric exists,
21
00:00:38,360 --> 00:00:40,520
you need to understand the problem it solves.
22
00:00:40,520 --> 00:00:42,600
20 years ago, if you were running a company,
23
00:00:42,600 --> 00:00:44,440
you'd buy a server for storing data,
24
00:00:44,440 --> 00:00:46,440
a separate tool for running reports,
25
00:00:46,440 --> 00:00:49,080
and yet another product for real-time analysis.
26
00:00:49,080 --> 00:00:51,600
Before you know it, you've got five different systems
27
00:00:51,600 --> 00:00:53,840
each with its own login, its own way of working,
28
00:00:53,840 --> 00:00:55,360
and its own team managing it.
29
00:00:55,360 --> 00:00:59,000
For a long time, data lived in silos, sales data set in one system,
30
00:00:59,000 --> 00:01:00,440
customer data in another,
31
00:01:00,440 --> 00:01:02,720
and financial data, somewhere else entirely.
32
00:01:02,720 --> 00:01:03,920
If you wanted to ask a question
33
00:01:03,920 --> 00:01:04,960
that crossed those systems,
34
00:01:04,960 --> 00:01:07,600
like which customers are buying the most profitable products,
35
00:01:07,600 --> 00:01:09,240
you had to move data around.
36
00:01:09,240 --> 00:01:11,920
Export it from one system, transform it, load it into another,
37
00:01:11,920 --> 00:01:13,640
then do it again when the data changed.
38
00:01:13,640 --> 00:01:16,120
More time was spent moving data between systems,
39
00:01:16,120 --> 00:01:17,360
than actually analyzing it.
40
00:01:17,360 --> 00:01:20,120
Data integration became a full-time job for entire teams.
41
00:01:20,120 --> 00:01:22,720
The worst part was that every time you moved data,
42
00:01:22,720 --> 00:01:25,640
you created another copy, another version of the truth,
43
00:01:25,640 --> 00:01:28,680
and pretty soon nobody knew which copy was the right one.
44
00:01:28,680 --> 00:01:30,480
Microsoft looked at this fragmentation
45
00:01:30,480 --> 00:01:31,800
and asked the simple question,
46
00:01:31,800 --> 00:01:33,520
"What if all these services work together?
47
00:01:33,520 --> 00:01:35,880
"What if instead of stitching together separate products,
48
00:01:35,880 --> 00:01:37,440
"you had one platform that did everything?"
49
00:01:37,440 --> 00:01:40,040
That question is what led to Microsoft fabric?
50
00:01:40,040 --> 00:01:42,840
What Microsoft fabric actually is?
51
00:01:42,840 --> 00:01:45,200
So what is fabric? Here's the simplest definition.
52
00:01:45,200 --> 00:01:48,680
Microsoft fabric is a unified data and analytics platform.
53
00:01:48,680 --> 00:01:50,200
One place for all your data work,
54
00:01:50,200 --> 00:01:52,080
think of it like a modern office building.
55
00:01:52,080 --> 00:01:54,360
Each tool like Power BI or Azure Synapse
56
00:01:54,360 --> 00:01:55,680
is a room inside that building.
57
00:01:55,680 --> 00:01:57,520
They have their own purpose, furniture, and tools,
58
00:01:57,520 --> 00:01:59,120
but they're all part of the same structure.
59
00:01:59,120 --> 00:02:01,760
You don't need to walk outside to get from one room to another.
60
00:02:01,760 --> 00:02:03,320
They're connected. That's what fabric does.
61
00:02:03,320 --> 00:02:04,840
It brings together data engineering,
62
00:02:04,840 --> 00:02:06,480
data integration, data warehousing,
63
00:02:06,480 --> 00:02:08,680
real-time analytics, and business intelligence,
64
00:02:08,680 --> 00:02:09,880
all in one platform.
65
00:02:09,880 --> 00:02:11,920
Everything runs on a single SaaS platform,
66
00:02:11,920 --> 00:02:14,280
which means no separate infrastructure to manage,
67
00:02:14,280 --> 00:02:16,760
no servers to provision on networking to configure.
68
00:02:16,760 --> 00:02:18,280
Microsoft handles all of that for you.
69
00:02:18,280 --> 00:02:19,840
Since its launch in 2023,
70
00:02:19,840 --> 00:02:22,040
fabric has become the fastest growing data product
71
00:02:22,040 --> 00:02:23,320
in Microsoft's history.
72
00:02:23,320 --> 00:02:25,440
Today, over 25,000 customers use it,
73
00:02:25,440 --> 00:02:27,720
including 80% of the Fortune 500.
74
00:02:27,720 --> 00:02:29,680
That's a lot of adoption in a very short time.
75
00:02:29,680 --> 00:02:30,960
The key difference is this.
76
00:02:30,960 --> 00:02:32,720
In the old world, you bought separate products
77
00:02:32,720 --> 00:02:34,720
and tried to make them talk to each other.
78
00:02:34,720 --> 00:02:37,360
With fabric, you don't stitch together different services anymore.
79
00:02:37,360 --> 00:02:39,840
It's one experience, one login, one way of working,
80
00:02:39,840 --> 00:02:41,560
and that changes everything.
81
00:02:41,560 --> 00:02:43,320
One lake, the foundation.
82
00:02:43,320 --> 00:02:44,560
Now let's talk about the foundation
83
00:02:44,560 --> 00:02:45,960
that makes all of this work.
84
00:02:45,960 --> 00:02:48,760
At the heart of fabric is something called one lake.
85
00:02:48,760 --> 00:02:50,400
The easiest way to think about one lake
86
00:02:50,400 --> 00:02:52,120
is like one drive for your data.
87
00:02:52,120 --> 00:02:53,800
You know how one drive gives you one place
88
00:02:53,800 --> 00:02:55,440
to store all your personal files
89
00:02:55,440 --> 00:02:58,080
and every app on your computer can access those files.
90
00:02:58,080 --> 00:02:59,560
One lake does the same thing,
91
00:02:59,560 --> 00:03:01,320
but for your organization's data.
92
00:03:01,320 --> 00:03:02,640
Here's what makes it powerful.
93
00:03:02,640 --> 00:03:04,040
Every fabric workload,
94
00:03:04,040 --> 00:03:05,560
whether it's a lake house, a warehouse,
95
00:03:05,560 --> 00:03:07,080
or a real-time analytics job,
96
00:03:07,080 --> 00:03:09,000
automatically stores its data in one lake.
97
00:03:09,000 --> 00:03:10,680
You don't have to manually copy anything
98
00:03:10,680 --> 00:03:12,160
or configure connections.
99
00:03:12,160 --> 00:03:13,160
It just happens.
100
00:03:13,160 --> 00:03:15,320
And there's only one one lake per tenant,
101
00:03:15,320 --> 00:03:18,200
one single source of truth for your entire organization.
102
00:03:18,200 --> 00:03:19,240
Now you might be thinking,
103
00:03:19,240 --> 00:03:21,000
what if I want to switch platforms later?
104
00:03:21,000 --> 00:03:22,080
Am I locked in?
105
00:03:22,080 --> 00:03:23,120
The answer is no.
106
00:03:23,120 --> 00:03:26,240
One lake stores data in open formats like Delta Parque,
107
00:03:26,240 --> 00:03:28,480
an industry standard format used by Databricks
108
00:03:28,480 --> 00:03:29,400
and Apache Spark.
109
00:03:29,400 --> 00:03:31,240
You're never locked into Microsoft.
110
00:03:31,240 --> 00:03:32,960
If you want to move your data somewhere else,
111
00:03:32,960 --> 00:03:33,480
you can.
112
00:03:33,480 --> 00:03:35,080
It's your data.
113
00:03:35,080 --> 00:03:37,000
But here's where one lake gets really interesting.
114
00:03:37,000 --> 00:03:37,960
Shortcuts.
115
00:03:37,960 --> 00:03:39,640
A shortcut is a virtual pointer to data
116
00:03:39,640 --> 00:03:40,640
that lives somewhere else.
117
00:03:40,640 --> 00:03:43,360
You can create a shortcut to data stored in AWS, S3,
118
00:03:43,360 --> 00:03:45,600
Google Cloud, or even on premises storage.
119
00:03:45,600 --> 00:03:48,160
And it shows up in one lake as if it was stored there.
120
00:03:48,160 --> 00:03:51,040
No data movement required, no duplication.
121
00:03:51,040 --> 00:03:53,360
You just point to it and it appears.
122
00:03:53,360 --> 00:03:55,920
One lake is built on top of Azure Data Lake Storage.
123
00:03:55,920 --> 00:03:58,080
Microsoft's enterprise-grade cloud storage.
124
00:03:58,080 --> 00:03:59,520
But you don't manage that infrastructure
125
00:03:59,520 --> 00:04:01,320
as it's all handled as a SaaS service.
126
00:04:01,320 --> 00:04:03,640
You don't provision servers or configure networking.
127
00:04:03,640 --> 00:04:04,400
You just use it.
128
00:04:04,400 --> 00:04:05,560
What does this mean for you?
129
00:04:05,560 --> 00:04:07,880
It means you can have data from multiple sources,
130
00:04:07,880 --> 00:04:10,160
cloud storage, on-premises databases,
131
00:04:10,160 --> 00:04:12,720
third-party platforms, all appear in one place
132
00:04:12,720 --> 00:04:13,880
without duplication.
133
00:04:13,880 --> 00:04:16,400
And the real power, any data written to one lake
134
00:04:16,400 --> 00:04:19,160
can be used by any fabric workload immediately.
135
00:04:19,160 --> 00:04:21,600
You write it once and every tool in fabric can see it.
136
00:04:21,600 --> 00:04:23,320
That's the magic of one lake.
137
00:04:23,320 --> 00:04:25,680
Lake houses, the data engineering hub.
138
00:04:25,680 --> 00:04:27,240
So you've got your data in one lake.
139
00:04:27,240 --> 00:04:28,360
Now you need to work with it.
140
00:04:28,360 --> 00:04:29,720
That's where the lake house comes in.
141
00:04:29,720 --> 00:04:30,960
Here's the simplest definition.
142
00:04:30,960 --> 00:04:32,920
A lake house is a data store that combines
143
00:04:32,920 --> 00:04:36,280
the flexibility of a data lake with the structure of a database.
144
00:04:36,280 --> 00:04:38,600
In a traditional data lake, you can store anything
145
00:04:38,600 --> 00:04:41,920
like the raw CSV files, JSON documents, images, video files,
146
00:04:41,920 --> 00:04:43,920
but it's messy with no structure or schema.
147
00:04:43,920 --> 00:04:45,680
So you can't query it easily.
148
00:04:45,680 --> 00:04:48,560
In a traditional database, everything is structured and queryable.
149
00:04:48,560 --> 00:04:50,560
But you can only store certain types of data.
150
00:04:50,560 --> 00:04:51,640
A lake house gives you both.
151
00:04:51,640 --> 00:04:55,240
You can keep raw files like CSV, JSON, or even images
152
00:04:55,240 --> 00:04:56,360
in a lake house.
153
00:04:56,360 --> 00:04:58,280
And you can also store structured tables
154
00:04:58,280 --> 00:04:59,680
that you can query with SQL.
155
00:04:59,680 --> 00:05:00,960
Both live in the same place.
156
00:05:00,960 --> 00:05:02,880
Those tables use Delta Park A format,
157
00:05:02,880 --> 00:05:05,040
the same open fast format used by Databricks
158
00:05:05,040 --> 00:05:06,520
and other major platforms designed
159
00:05:06,520 --> 00:05:08,360
to handle massive amounts of data.
160
00:05:08,360 --> 00:05:10,280
Lake houses support multiple tools.
161
00:05:10,280 --> 00:05:12,960
You can use Spark for large-scale data processing.
162
00:05:12,960 --> 00:05:14,560
You can use Python for machine learning.
163
00:05:14,560 --> 00:05:16,440
You can use SQL for simple queries.
164
00:05:16,440 --> 00:05:18,160
Whatever tool you prefer, the lake house
165
00:05:18,160 --> 00:05:19,280
can handle it.
166
00:05:19,280 --> 00:05:21,040
Think of a lake house as your workspace
167
00:05:21,040 --> 00:05:22,640
for building data pipelines.
168
00:05:22,640 --> 00:05:25,400
You bring raw data in, transform it, clean it,
169
00:05:25,400 --> 00:05:26,680
and prepared for analysis.
170
00:05:26,680 --> 00:05:28,720
You can run machine learning models on it.
171
00:05:28,720 --> 00:05:30,840
You can create training datasets all in one place.
172
00:05:30,840 --> 00:05:33,440
For data engineers who prefer a code-friendly environment,
173
00:05:33,440 --> 00:05:35,160
lake houses support notebooks.
174
00:05:35,160 --> 00:05:37,120
These are interactive documents where you write code,
175
00:05:37,120 --> 00:05:38,880
see results, and document your work,
176
00:05:38,880 --> 00:05:40,520
like a lab notebook for data.
177
00:05:40,520 --> 00:05:42,200
You can write Spark code in one cell,
178
00:05:42,200 --> 00:05:44,440
see the output in the next, and explain what you did
179
00:05:44,440 --> 00:05:45,360
in the cell after that.
180
00:05:45,360 --> 00:05:46,600
The lake house is the entry point
181
00:05:46,600 --> 00:05:49,040
for most data engineers working in fabric.
182
00:05:49,040 --> 00:05:50,760
It's where the heavy lifting happens.
183
00:05:50,760 --> 00:05:52,480
And because it stores everything in one lake,
184
00:05:52,480 --> 00:05:55,160
any data you put in a lake house is immediately available
185
00:05:55,160 --> 00:05:58,240
to every other fabric workload in the warehouses,
186
00:05:58,240 --> 00:05:59,760
the SQL Analytics engine.
187
00:05:59,760 --> 00:06:01,800
Now, if the lake house is the data engineering hub,
188
00:06:01,800 --> 00:06:04,120
think of the warehouse as the SQL Analytics engine,
189
00:06:04,120 --> 00:06:05,880
built for a different kind of user.
190
00:06:05,880 --> 00:06:07,920
A warehouse in fabric is a fully managed
191
00:06:07,920 --> 00:06:09,680
SQL Analytics database.
192
00:06:09,680 --> 00:06:12,240
If you're a database administrator or a SQL developer,
193
00:06:12,240 --> 00:06:13,840
this is where you'll feel at home.
194
00:06:13,840 --> 00:06:15,200
You can write T-Sycle queries.
195
00:06:15,200 --> 00:06:18,600
You can create views, stored procedures, and security policies.
196
00:06:18,600 --> 00:06:20,600
It looks and feels like a traditional data warehouse,
197
00:06:20,600 --> 00:06:22,040
but here's what makes it different.
198
00:06:22,040 --> 00:06:23,440
Unlike a traditional data warehouse,
199
00:06:23,440 --> 00:06:26,120
fabrics warehouse stores data directly in one lake.
200
00:06:26,120 --> 00:06:27,880
There's no separate storage layer,
201
00:06:27,880 --> 00:06:29,400
no extra infrastructure to manage.
202
00:06:29,400 --> 00:06:30,800
The data lives in one lake,
203
00:06:30,800 --> 00:06:33,640
and the warehouse is just the interface you use to query it.
204
00:06:33,640 --> 00:06:36,160
This matters because the warehouse and the lake house
205
00:06:36,160 --> 00:06:38,640
share the same data, so you're not duplicating anything.
206
00:06:38,640 --> 00:06:41,080
If a data engineer loads data into a lake house,
207
00:06:41,080 --> 00:06:43,760
a SQL developer can query it from a warehouse immediately,
208
00:06:43,760 --> 00:06:46,720
no copying, no moving, it's the same data accessed
209
00:06:46,720 --> 00:06:48,040
through different tools.
210
00:06:48,040 --> 00:06:51,000
The warehouse is built for large-scale analytics.
211
00:06:51,000 --> 00:06:53,200
Under the hood, it uses massive parallel processing
212
00:06:53,200 --> 00:06:54,960
to handle huge data sets.
213
00:06:54,960 --> 00:06:57,320
And Microsoft has been investing heavily in performance.
214
00:06:57,320 --> 00:06:58,920
In the last six months alone,
215
00:06:58,920 --> 00:07:01,520
performance has improved by 36%.
216
00:07:01,520 --> 00:07:03,200
That's a lot of improvement in a short time.
217
00:07:03,200 --> 00:07:04,920
If you're migrating from another platform,
218
00:07:04,920 --> 00:07:06,600
fabric has a migration assistant
219
00:07:06,600 --> 00:07:08,800
that helps you move from snowflake or synapse.
220
00:07:08,800 --> 00:07:10,320
It analyzes your existing setup
221
00:07:10,320 --> 00:07:11,800
and guides you through the process.
222
00:07:11,800 --> 00:07:14,200
The goal is to make migration as painless as possible.
223
00:07:14,200 --> 00:07:16,400
So what's the difference between a lake house and a warehouse?
224
00:07:16,400 --> 00:07:17,520
It comes down to the interface.
225
00:07:17,520 --> 00:07:19,680
We're houses are SQL first.
226
00:07:19,680 --> 00:07:21,640
They're built for people who think in terms of tables,
227
00:07:21,640 --> 00:07:23,480
queries and stored procedures.
228
00:07:23,480 --> 00:07:25,280
Lake houses are developer first.
229
00:07:25,280 --> 00:07:27,600
They're built for people who want to use Spark, Python,
230
00:07:27,600 --> 00:07:28,520
and notebooks.
231
00:07:28,520 --> 00:07:30,960
But underneath, they both store data in one lake.
232
00:07:30,960 --> 00:07:33,160
There are two ways of working with the same data.
233
00:07:33,160 --> 00:07:36,320
Getting data in, data factory, mirroring, and shortcuts.
234
00:07:36,320 --> 00:07:37,960
So you've got one lake for storage,
235
00:07:37,960 --> 00:07:40,720
lake houses for transformation and warehouses for querying.
236
00:07:40,720 --> 00:07:42,720
But how does the data actually get in?
237
00:07:42,720 --> 00:07:44,120
That's where data factory steps in.
238
00:07:44,120 --> 00:07:47,080
Data factory is Fabrics built in data integration service.
239
00:07:47,080 --> 00:07:48,560
And it gives you over 200 connectors
240
00:07:48,560 --> 00:07:51,880
to pull data from databases, cloud storage, SAS apps, and more.
241
00:07:51,880 --> 00:07:54,240
It's the most used data integration tool in the world
242
00:07:54,240 --> 00:07:55,400
and for good reason.
243
00:07:55,400 --> 00:07:57,920
Within data factory, you get data flows gen 2,
244
00:07:57,920 --> 00:08:00,040
which gives you a no-code way to transform data
245
00:08:00,040 --> 00:08:01,200
using power query.
246
00:08:01,200 --> 00:08:03,640
If you've ever used power query in Excel or Power BI,
247
00:08:03,640 --> 00:08:04,800
you already know the drill.
248
00:08:04,800 --> 00:08:06,240
You click through a visual interface
249
00:08:06,240 --> 00:08:08,840
to clean reshape and combine data, no coding required,
250
00:08:08,840 --> 00:08:11,520
and it's enterprise-grade, scalable to handle massive data
251
00:08:11,520 --> 00:08:12,240
sets.
252
00:08:12,240 --> 00:08:15,920
Over 22 billion orchestrations run on data factory every month,
253
00:08:15,920 --> 00:08:18,640
which means billions of data movements happen automatically,
254
00:08:18,640 --> 00:08:20,760
making it a serious piece of infrastructure.
255
00:08:20,760 --> 00:08:24,280
But there's an even simpler way to get data into one lake, mirroring.
256
00:08:24,280 --> 00:08:26,560
Mirroring copies data from external databases
257
00:08:26,560 --> 00:08:29,200
like Azure, School, Oracle, or Google BigQuery
258
00:08:29,200 --> 00:08:30,760
into one lake in near real time.
259
00:08:30,760 --> 00:08:32,280
And here's the best part.
260
00:08:32,280 --> 00:08:35,280
It's free, included in the service with no extra cost.
261
00:08:35,280 --> 00:08:36,800
You just point to your source database
262
00:08:36,800 --> 00:08:38,360
and Fabric handles the rest.
263
00:08:38,360 --> 00:08:40,560
Then there are shortcuts, which we touched on earlier,
264
00:08:40,560 --> 00:08:42,160
but they deserve more attention.
265
00:08:42,160 --> 00:08:44,840
Shortcuts are virtual pointers to data stored elsewhere,
266
00:08:44,840 --> 00:08:46,360
requiring no data movement.
267
00:08:46,360 --> 00:08:48,960
You create a shortcut to an AWS tree bucket,
268
00:08:48,960 --> 00:08:51,440
and it shows up in one lake as if it was stored there,
269
00:08:51,440 --> 00:08:52,920
while the data stays where it is.
270
00:08:52,920 --> 00:08:54,200
You're just pointing to it.
271
00:08:54,200 --> 00:08:56,880
And now shortcuts can do more than just point.
272
00:08:56,880 --> 00:08:59,840
Shortcut transforms let you apply AI-powered transformations
273
00:08:59,840 --> 00:09:00,680
on the fly.
274
00:09:00,680 --> 00:09:04,160
You can run sentiment analysis on text data, translate content,
275
00:09:04,160 --> 00:09:06,560
or detect personally identifiable information
276
00:09:06,560 --> 00:09:09,200
all without moving the data or building a pipeline.
277
00:09:09,200 --> 00:09:11,640
The transformation happens as you access it.
278
00:09:11,640 --> 00:09:13,600
The goal of all these tools is simple.
279
00:09:13,600 --> 00:09:16,360
Bring all your data into one lake without building complex
280
00:09:16,360 --> 00:09:19,520
pipelines, whether you use data factory for traditional ETL,
281
00:09:19,520 --> 00:09:21,480
mirroring for near real time replication
282
00:09:21,480 --> 00:09:24,840
or shortcuts for virtual access, the result is the same.
283
00:09:24,840 --> 00:09:29,120
Your data ends up in one lake ready for any Fabric workload.
284
00:09:29,120 --> 00:09:31,440
Real-time intelligence databases in AI.
285
00:09:31,440 --> 00:09:33,640
So far, we've talked about moving data and storing it,
286
00:09:33,640 --> 00:09:35,720
but what about data that never stops moving?
287
00:09:35,720 --> 00:09:38,640
Streaming data from IoT sensors, website click streams,
288
00:09:38,640 --> 00:09:41,520
telemetry from devices, that's where real-time intelligence
289
00:09:41,520 --> 00:09:42,240
comes in.
290
00:09:42,240 --> 00:09:44,440
Fabric handles streaming data through something called
291
00:09:44,440 --> 00:09:46,640
an event house, which stores and queries
292
00:09:46,640 --> 00:09:49,160
high volume time series data using KQL,
293
00:09:49,160 --> 00:09:50,720
the custochery language.
294
00:09:50,720 --> 00:09:53,000
It's the same engine that powers Azure Data Explorer
295
00:09:53,000 --> 00:09:56,000
built for speed, so you can ingest millions of events per second
296
00:09:56,000 --> 00:09:57,520
and query them in milliseconds.
297
00:09:57,520 --> 00:09:58,280
This matters.
298
00:09:58,280 --> 00:09:59,160
And here's why.
299
00:09:59,160 --> 00:10:02,520
46% of fabric customers already use real-time intelligence.
300
00:10:02,520 --> 00:10:03,600
That's nearly half.
301
00:10:03,600 --> 00:10:05,240
The use cases are everywhere.
302
00:10:05,240 --> 00:10:07,440
A logistics company tracking delivery trucks,
303
00:10:07,440 --> 00:10:09,480
a retailer monitoring in store for traffic,
304
00:10:09,480 --> 00:10:12,360
a manufacturer watching sensor data from factory equipment,
305
00:10:12,360 --> 00:10:14,200
all happening in real time.
306
00:10:14,200 --> 00:10:15,960
But fabric isn't just about analytics.
307
00:10:15,960 --> 00:10:17,680
It also includes operational databases
308
00:10:17,680 --> 00:10:19,440
like a SQL database for relational data
309
00:10:19,440 --> 00:10:21,440
and Cosmos DB for no-school workloads.
310
00:10:21,440 --> 00:10:24,520
These are the same databases you'd use to run your applications.
311
00:10:24,520 --> 00:10:26,120
And because they're built into Fabric,
312
00:10:26,120 --> 00:10:28,760
they automatically store their data in one lake.
313
00:10:28,760 --> 00:10:30,800
No extra pipelines, no manual copying,
314
00:10:30,800 --> 00:10:34,200
your transactional data and analytical data live in the same place.
315
00:10:34,200 --> 00:10:36,400
Now let's talk about AI because this is where fabric
316
00:10:36,400 --> 00:10:37,680
gets really interesting.
317
00:10:37,680 --> 00:10:40,000
AI integration runs deep across the platform
318
00:10:40,000 --> 00:10:42,120
and you get pre-built models for text analysis,
319
00:10:42,120 --> 00:10:43,880
translation and sentiment analysis.
320
00:10:43,880 --> 00:10:45,600
You can use them directly in your data flows
321
00:10:45,600 --> 00:10:48,160
without writing any code, just point to your data,
322
00:10:48,160 --> 00:10:49,680
select the model and it runs.
323
00:10:49,680 --> 00:10:52,440
And then there's Copilot, which is embedded across fabric.
324
00:10:52,440 --> 00:10:54,320
It helps you write queries, build reports,
325
00:10:54,320 --> 00:10:55,360
and create pipelines.
326
00:10:55,360 --> 00:10:57,400
You just describe what you want in plain English
327
00:10:57,400 --> 00:10:59,080
and Copilot generates the code.
328
00:10:59,080 --> 00:11:00,560
It's like having a junior data engineer
329
00:11:00,560 --> 00:11:02,960
sitting next to you ready to help at any moment.
330
00:11:02,960 --> 00:11:04,240
The vision here is simple.
331
00:11:04,240 --> 00:11:06,680
AI works on your data without moving it anywhere.
332
00:11:06,680 --> 00:11:09,640
Your data stays in one lake and the AI models come to the data.
333
00:11:09,640 --> 00:11:12,920
No copying, no exporting, no security risks.
334
00:11:12,920 --> 00:11:15,120
The data never leaves your governed environment.
335
00:11:15,120 --> 00:11:18,960
Power BI, governance and the business value.
336
00:11:18,960 --> 00:11:20,800
Now we get to the part most people actually see
337
00:11:20,800 --> 00:11:23,000
the front door to Insights Power BI.
338
00:11:23,000 --> 00:11:25,560
It's used by 550,000 organizations
339
00:11:25,560 --> 00:11:27,960
and 34 million users every month,
340
00:11:27,960 --> 00:11:30,680
making it the most widely used business intelligence tool
341
00:11:30,680 --> 00:11:31,640
in the world.
342
00:11:31,640 --> 00:11:34,120
And in fabric, it becomes something even more powerful.
343
00:11:34,120 --> 00:11:35,040
Here's the key.
344
00:11:35,040 --> 00:11:37,960
In fabric, Power BI connects directly to one lake
345
00:11:37,960 --> 00:11:39,920
using something called direct lake mode.
346
00:11:39,920 --> 00:11:41,640
This is different from the traditional import mode
347
00:11:41,640 --> 00:11:44,440
where you copy data into Power BI's internal storage.
348
00:11:44,440 --> 00:11:46,120
Instead, direct lake queries the data
349
00:11:46,120 --> 00:11:48,520
where it sits in one lake with no refresh needed,
350
00:11:48,520 --> 00:11:50,440
no duplication and the data is always current
351
00:11:50,440 --> 00:11:51,840
and blazing fast.
352
00:11:51,840 --> 00:11:55,120
Underneath every Power BI report is a semantic model.
353
00:11:55,120 --> 00:11:57,680
The brain of your report where you define business logic,
354
00:11:57,680 --> 00:11:59,360
measures and relationships.
355
00:11:59,360 --> 00:12:01,440
And now you can build these models entirely
356
00:12:01,440 --> 00:12:02,560
in a web browser.
357
00:12:02,560 --> 00:12:04,040
Web modeling is generally available
358
00:12:04,040 --> 00:12:06,720
so you can create a semantic model, build relationships
359
00:12:06,720 --> 00:12:09,560
and write DAX measures all from a browser even on a Mac.
360
00:12:09,560 --> 00:12:11,800
No desktop app required, but all this data power
361
00:12:11,800 --> 00:12:13,320
means nothing without governance.
362
00:12:13,320 --> 00:12:15,680
And fabric has governance built in from the start.
363
00:12:15,680 --> 00:12:17,960
Microsoft Perview provides sensitivity labels
364
00:12:17,960 --> 00:12:19,480
that follow your data everywhere.
365
00:12:19,480 --> 00:12:21,640
If a report contains sensitive information
366
00:12:21,640 --> 00:12:24,080
that labels stays with it, even when exported.
367
00:12:24,080 --> 00:12:26,680
Data loss prevention policies automatically detect
368
00:12:26,680 --> 00:12:29,800
sensitive data and block unauthorized sharing
369
00:12:29,800 --> 00:12:33,800
and access controls ensure only the right people see the right data.
370
00:12:33,800 --> 00:12:36,000
Domains help you organize data by business area
371
00:12:36,000 --> 00:12:37,480
with separate governance rules
372
00:12:37,480 --> 00:12:39,600
so your finance data can have stricter controls
373
00:12:39,600 --> 00:12:41,160
than your marketing data.
374
00:12:41,160 --> 00:12:43,120
Each domain can be managed by its own team,
375
00:12:43,120 --> 00:12:45,920
making it governance that scales with your organization.
376
00:12:45,920 --> 00:12:47,680
Now let's talk about real results.
377
00:12:47,680 --> 00:12:48,920
This isn't just theory.
378
00:12:48,920 --> 00:12:51,720
Eastman Chemical used fabric to reduce sales preparation
379
00:12:51,720 --> 00:12:54,960
time by 83%, what used to take four hours now
380
00:12:54,960 --> 00:12:56,400
takes about 40 minutes.
381
00:12:56,400 --> 00:12:59,240
Eaton Solutions cut manual effort by 75%
382
00:12:59,240 --> 00:13:02,520
with processes that require 10 to 15 steps now taking two.
383
00:13:02,520 --> 00:13:04,680
And Sonata Software saves 30,000 hours
384
00:13:04,680 --> 00:13:06,560
of manual reconciliation every year,
385
00:13:06,560 --> 00:13:09,240
redirecting that human effort to higher value work.
386
00:13:09,240 --> 00:13:10,960
The cost model is straightforward.
387
00:13:10,960 --> 00:13:14,200
You pay for capacity, the compute power that runs your workloads.
388
00:13:14,200 --> 00:13:16,400
Microsoft has built in smoothing and search protection
389
00:13:16,400 --> 00:13:17,480
to manage peaks.
390
00:13:17,480 --> 00:13:19,560
Smoothing spreads the cost of background operations
391
00:13:19,560 --> 00:13:22,480
over 24 hours, so you don't get hit with sudden spikes
392
00:13:22,480 --> 00:13:24,560
and search protection limits background jobs
393
00:13:24,560 --> 00:13:27,520
during busy periods to keep interactive performance stable.
394
00:13:27,520 --> 00:13:28,880
You pay for what you use
395
00:13:28,880 --> 00:13:31,080
and you don't get surprised.
396
00:13:31,080 --> 00:13:32,320
Let's bring this all together.
397
00:13:32,320 --> 00:13:34,640
Fabric isn't just another Microsoft product.
398
00:13:34,640 --> 00:13:37,520
It's a complete rethinking of how data platforms should work.
399
00:13:37,520 --> 00:13:41,040
One platform, one lake, one copy of data and infinite possibilities.
400
00:13:41,040 --> 00:13:43,760
And with AI and co-pilot built in, it's accessible to everyone,
401
00:13:43,760 --> 00:13:46,480
not just data professionals, but business analysts, sales teams,
402
00:13:46,480 --> 00:13:48,880
operations managers, anyone who needs answers from data.
403
00:13:48,880 --> 00:13:49,880
Here's your homework.
404
00:13:49,880 --> 00:13:52,960
Go to fabric, Microsoft.com, start a free trial
405
00:13:52,960 --> 00:13:54,480
and create your first workspace.
406
00:13:54,480 --> 00:13:55,280
Just try it.
407
00:13:55,280 --> 00:13:57,440
You'll see how the pieces fit together in a way
408
00:13:57,440 --> 00:13:59,440
that no amount of reading can replace.
409
00:13:59,440 --> 00:14:01,560
If this episode helped you understand fabric better,
410
00:14:01,560 --> 00:14:03,320
subscribe on your favorite podcast platform
411
00:14:03,320 --> 00:14:05,600
and share it with someone who's starting their data journey.
412
00:14:05,600 --> 00:14:06,520
They'll thank you.
413
00:14:06,520 --> 00:14:10,320
Next time, we'll break down one lake short cuts in plain English.
414
00:14:10,320 --> 00:14:12,200
How they work, why they matter,
415
00:14:12,200 --> 00:14:14,840
and how to use them without moving data around.
416
00:14:14,840 --> 00:14:16,080
That's one you won't want to miss.
417
00:14:16,080 --> 00:14:17,480
Until then, keep learning.
00:00:00,000 --> 00:00:01,520
Welcome to another episode.
2
00:00:01,520 --> 00:00:03,360
Today's topic is one you've probably heard
3
00:00:03,360 --> 00:00:05,000
in meetings or tech articles.
4
00:00:05,000 --> 00:00:07,480
Microsoft fabric, but if you're like most people,
5
00:00:07,480 --> 00:00:09,680
you're not entirely sure what it actually is.
6
00:00:09,680 --> 00:00:11,360
Is it just another Microsoft product,
7
00:00:11,360 --> 00:00:12,600
a rebrand of something old,
8
00:00:12,600 --> 00:00:14,320
or is it something completely different?
9
00:00:14,320 --> 00:00:15,800
Let me give you the short answer.
10
00:00:15,800 --> 00:00:17,760
Microsoft fabric is not just another product.
11
00:00:17,760 --> 00:00:19,040
It's a complete data platform,
12
00:00:19,040 --> 00:00:21,240
one place where you can do all your data work,
13
00:00:21,240 --> 00:00:24,400
storage, transformation, analytics, and reporting,
14
00:00:24,400 --> 00:00:25,440
all in one place.
15
00:00:25,440 --> 00:00:26,680
By the end of this episode,
16
00:00:26,680 --> 00:00:29,400
you'll understand what fabric is, why Microsoft built it,
17
00:00:29,400 --> 00:00:31,240
and how the main pieces fit together.
18
00:00:31,240 --> 00:00:33,440
Grab your coffee and let's dive in.
19
00:00:33,440 --> 00:00:36,320
The old way, why data was broken?
20
00:00:36,320 --> 00:00:38,360
To understand why fabric exists,
21
00:00:38,360 --> 00:00:40,520
you need to understand the problem it solves.
22
00:00:40,520 --> 00:00:42,600
20 years ago, if you were running a company,
23
00:00:42,600 --> 00:00:44,440
you'd buy a server for storing data,
24
00:00:44,440 --> 00:00:46,440
a separate tool for running reports,
25
00:00:46,440 --> 00:00:49,080
and yet another product for real-time analysis.
26
00:00:49,080 --> 00:00:51,600
Before you know it, you've got five different systems
27
00:00:51,600 --> 00:00:53,840
each with its own login, its own way of working,
28
00:00:53,840 --> 00:00:55,360
and its own team managing it.
29
00:00:55,360 --> 00:00:59,000
For a long time, data lived in silos, sales data set in one system,
30
00:00:59,000 --> 00:01:00,440
customer data in another,
31
00:01:00,440 --> 00:01:02,720
and financial data, somewhere else entirely.
32
00:01:02,720 --> 00:01:03,920
If you wanted to ask a question
33
00:01:03,920 --> 00:01:04,960
that crossed those systems,
34
00:01:04,960 --> 00:01:07,600
like which customers are buying the most profitable products,
35
00:01:07,600 --> 00:01:09,240
you had to move data around.
36
00:01:09,240 --> 00:01:11,920
Export it from one system, transform it, load it into another,
37
00:01:11,920 --> 00:01:13,640
then do it again when the data changed.
38
00:01:13,640 --> 00:01:16,120
More time was spent moving data between systems,
39
00:01:16,120 --> 00:01:17,360
than actually analyzing it.
40
00:01:17,360 --> 00:01:20,120
Data integration became a full-time job for entire teams.
41
00:01:20,120 --> 00:01:22,720
The worst part was that every time you moved data,
42
00:01:22,720 --> 00:01:25,640
you created another copy, another version of the truth,
43
00:01:25,640 --> 00:01:28,680
and pretty soon nobody knew which copy was the right one.
44
00:01:28,680 --> 00:01:30,480
Microsoft looked at this fragmentation
45
00:01:30,480 --> 00:01:31,800
and asked the simple question,
46
00:01:31,800 --> 00:01:33,520
"What if all these services work together?
47
00:01:33,520 --> 00:01:35,880
"What if instead of stitching together separate products,
48
00:01:35,880 --> 00:01:37,440
"you had one platform that did everything?"
49
00:01:37,440 --> 00:01:40,040
That question is what led to Microsoft fabric?
50
00:01:40,040 --> 00:01:42,840
What Microsoft fabric actually is?
51
00:01:42,840 --> 00:01:45,200
So what is fabric? Here's the simplest definition.
52
00:01:45,200 --> 00:01:48,680
Microsoft fabric is a unified data and analytics platform.
53
00:01:48,680 --> 00:01:50,200
One place for all your data work,
54
00:01:50,200 --> 00:01:52,080
think of it like a modern office building.
55
00:01:52,080 --> 00:01:54,360
Each tool like Power BI or Azure Synapse
56
00:01:54,360 --> 00:01:55,680
is a room inside that building.
57
00:01:55,680 --> 00:01:57,520
They have their own purpose, furniture, and tools,
58
00:01:57,520 --> 00:01:59,120
but they're all part of the same structure.
59
00:01:59,120 --> 00:02:01,760
You don't need to walk outside to get from one room to another.
60
00:02:01,760 --> 00:02:03,320
They're connected. That's what fabric does.
61
00:02:03,320 --> 00:02:04,840
It brings together data engineering,
62
00:02:04,840 --> 00:02:06,480
data integration, data warehousing,
63
00:02:06,480 --> 00:02:08,680
real-time analytics, and business intelligence,
64
00:02:08,680 --> 00:02:09,880
all in one platform.
65
00:02:09,880 --> 00:02:11,920
Everything runs on a single SaaS platform,
66
00:02:11,920 --> 00:02:14,280
which means no separate infrastructure to manage,
67
00:02:14,280 --> 00:02:16,760
no servers to provision on networking to configure.
68
00:02:16,760 --> 00:02:18,280
Microsoft handles all of that for you.
69
00:02:18,280 --> 00:02:19,840
Since its launch in 2023,
70
00:02:19,840 --> 00:02:22,040
fabric has become the fastest growing data product
71
00:02:22,040 --> 00:02:23,320
in Microsoft's history.
72
00:02:23,320 --> 00:02:25,440
Today, over 25,000 customers use it,
73
00:02:25,440 --> 00:02:27,720
including 80% of the Fortune 500.
74
00:02:27,720 --> 00:02:29,680
That's a lot of adoption in a very short time.
75
00:02:29,680 --> 00:02:30,960
The key difference is this.
76
00:02:30,960 --> 00:02:32,720
In the old world, you bought separate products
77
00:02:32,720 --> 00:02:34,720
and tried to make them talk to each other.
78
00:02:34,720 --> 00:02:37,360
With fabric, you don't stitch together different services anymore.
79
00:02:37,360 --> 00:02:39,840
It's one experience, one login, one way of working,
80
00:02:39,840 --> 00:02:41,560
and that changes everything.
81
00:02:41,560 --> 00:02:43,320
One lake, the foundation.
82
00:02:43,320 --> 00:02:44,560
Now let's talk about the foundation
83
00:02:44,560 --> 00:02:45,960
that makes all of this work.
84
00:02:45,960 --> 00:02:48,760
At the heart of fabric is something called one lake.
85
00:02:48,760 --> 00:02:50,400
The easiest way to think about one lake
86
00:02:50,400 --> 00:02:52,120
is like one drive for your data.
87
00:02:52,120 --> 00:02:53,800
You know how one drive gives you one place
88
00:02:53,800 --> 00:02:55,440
to store all your personal files
89
00:02:55,440 --> 00:02:58,080
and every app on your computer can access those files.
90
00:02:58,080 --> 00:02:59,560
One lake does the same thing,
91
00:02:59,560 --> 00:03:01,320
but for your organization's data.
92
00:03:01,320 --> 00:03:02,640
Here's what makes it powerful.
93
00:03:02,640 --> 00:03:04,040
Every fabric workload,
94
00:03:04,040 --> 00:03:05,560
whether it's a lake house, a warehouse,
95
00:03:05,560 --> 00:03:07,080
or a real-time analytics job,
96
00:03:07,080 --> 00:03:09,000
automatically stores its data in one lake.
97
00:03:09,000 --> 00:03:10,680
You don't have to manually copy anything
98
00:03:10,680 --> 00:03:12,160
or configure connections.
99
00:03:12,160 --> 00:03:13,160
It just happens.
100
00:03:13,160 --> 00:03:15,320
And there's only one one lake per tenant,
101
00:03:15,320 --> 00:03:18,200
one single source of truth for your entire organization.
102
00:03:18,200 --> 00:03:19,240
Now you might be thinking,
103
00:03:19,240 --> 00:03:21,000
what if I want to switch platforms later?
104
00:03:21,000 --> 00:03:22,080
Am I locked in?
105
00:03:22,080 --> 00:03:23,120
The answer is no.
106
00:03:23,120 --> 00:03:26,240
One lake stores data in open formats like Delta Parque,
107
00:03:26,240 --> 00:03:28,480
an industry standard format used by Databricks
108
00:03:28,480 --> 00:03:29,400
and Apache Spark.
109
00:03:29,400 --> 00:03:31,240
You're never locked into Microsoft.
110
00:03:31,240 --> 00:03:32,960
If you want to move your data somewhere else,
111
00:03:32,960 --> 00:03:33,480
you can.
112
00:03:33,480 --> 00:03:35,080
It's your data.
113
00:03:35,080 --> 00:03:37,000
But here's where one lake gets really interesting.
114
00:03:37,000 --> 00:03:37,960
Shortcuts.
115
00:03:37,960 --> 00:03:39,640
A shortcut is a virtual pointer to data
116
00:03:39,640 --> 00:03:40,640
that lives somewhere else.
117
00:03:40,640 --> 00:03:43,360
You can create a shortcut to data stored in AWS, S3,
118
00:03:43,360 --> 00:03:45,600
Google Cloud, or even on premises storage.
119
00:03:45,600 --> 00:03:48,160
And it shows up in one lake as if it was stored there.
120
00:03:48,160 --> 00:03:51,040
No data movement required, no duplication.
121
00:03:51,040 --> 00:03:53,360
You just point to it and it appears.
122
00:03:53,360 --> 00:03:55,920
One lake is built on top of Azure Data Lake Storage.
123
00:03:55,920 --> 00:03:58,080
Microsoft's enterprise-grade cloud storage.
124
00:03:58,080 --> 00:03:59,520
But you don't manage that infrastructure
125
00:03:59,520 --> 00:04:01,320
as it's all handled as a SaaS service.
126
00:04:01,320 --> 00:04:03,640
You don't provision servers or configure networking.
127
00:04:03,640 --> 00:04:04,400
You just use it.
128
00:04:04,400 --> 00:04:05,560
What does this mean for you?
129
00:04:05,560 --> 00:04:07,880
It means you can have data from multiple sources,
130
00:04:07,880 --> 00:04:10,160
cloud storage, on-premises databases,
131
00:04:10,160 --> 00:04:12,720
third-party platforms, all appear in one place
132
00:04:12,720 --> 00:04:13,880
without duplication.
133
00:04:13,880 --> 00:04:16,400
And the real power, any data written to one lake
134
00:04:16,400 --> 00:04:19,160
can be used by any fabric workload immediately.
135
00:04:19,160 --> 00:04:21,600
You write it once and every tool in fabric can see it.
136
00:04:21,600 --> 00:04:23,320
That's the magic of one lake.
137
00:04:23,320 --> 00:04:25,680
Lake houses, the data engineering hub.
138
00:04:25,680 --> 00:04:27,240
So you've got your data in one lake.
139
00:04:27,240 --> 00:04:28,360
Now you need to work with it.
140
00:04:28,360 --> 00:04:29,720
That's where the lake house comes in.
141
00:04:29,720 --> 00:04:30,960
Here's the simplest definition.
142
00:04:30,960 --> 00:04:32,920
A lake house is a data store that combines
143
00:04:32,920 --> 00:04:36,280
the flexibility of a data lake with the structure of a database.
144
00:04:36,280 --> 00:04:38,600
In a traditional data lake, you can store anything
145
00:04:38,600 --> 00:04:41,920
like the raw CSV files, JSON documents, images, video files,
146
00:04:41,920 --> 00:04:43,920
but it's messy with no structure or schema.
147
00:04:43,920 --> 00:04:45,680
So you can't query it easily.
148
00:04:45,680 --> 00:04:48,560
In a traditional database, everything is structured and queryable.
149
00:04:48,560 --> 00:04:50,560
But you can only store certain types of data.
150
00:04:50,560 --> 00:04:51,640
A lake house gives you both.
151
00:04:51,640 --> 00:04:55,240
You can keep raw files like CSV, JSON, or even images
152
00:04:55,240 --> 00:04:56,360
in a lake house.
153
00:04:56,360 --> 00:04:58,280
And you can also store structured tables
154
00:04:58,280 --> 00:04:59,680
that you can query with SQL.
155
00:04:59,680 --> 00:05:00,960
Both live in the same place.
156
00:05:00,960 --> 00:05:02,880
Those tables use Delta Park A format,
157
00:05:02,880 --> 00:05:05,040
the same open fast format used by Databricks
158
00:05:05,040 --> 00:05:06,520
and other major platforms designed
159
00:05:06,520 --> 00:05:08,360
to handle massive amounts of data.
160
00:05:08,360 --> 00:05:10,280
Lake houses support multiple tools.
161
00:05:10,280 --> 00:05:12,960
You can use Spark for large-scale data processing.
162
00:05:12,960 --> 00:05:14,560
You can use Python for machine learning.
163
00:05:14,560 --> 00:05:16,440
You can use SQL for simple queries.
164
00:05:16,440 --> 00:05:18,160
Whatever tool you prefer, the lake house
165
00:05:18,160 --> 00:05:19,280
can handle it.
166
00:05:19,280 --> 00:05:21,040
Think of a lake house as your workspace
167
00:05:21,040 --> 00:05:22,640
for building data pipelines.
168
00:05:22,640 --> 00:05:25,400
You bring raw data in, transform it, clean it,
169
00:05:25,400 --> 00:05:26,680
and prepared for analysis.
170
00:05:26,680 --> 00:05:28,720
You can run machine learning models on it.
171
00:05:28,720 --> 00:05:30,840
You can create training datasets all in one place.
172
00:05:30,840 --> 00:05:33,440
For data engineers who prefer a code-friendly environment,
173
00:05:33,440 --> 00:05:35,160
lake houses support notebooks.
174
00:05:35,160 --> 00:05:37,120
These are interactive documents where you write code,
175
00:05:37,120 --> 00:05:38,880
see results, and document your work,
176
00:05:38,880 --> 00:05:40,520
like a lab notebook for data.
177
00:05:40,520 --> 00:05:42,200
You can write Spark code in one cell,
178
00:05:42,200 --> 00:05:44,440
see the output in the next, and explain what you did
179
00:05:44,440 --> 00:05:45,360
in the cell after that.
180
00:05:45,360 --> 00:05:46,600
The lake house is the entry point
181
00:05:46,600 --> 00:05:49,040
for most data engineers working in fabric.
182
00:05:49,040 --> 00:05:50,760
It's where the heavy lifting happens.
183
00:05:50,760 --> 00:05:52,480
And because it stores everything in one lake,
184
00:05:52,480 --> 00:05:55,160
any data you put in a lake house is immediately available
185
00:05:55,160 --> 00:05:58,240
to every other fabric workload in the warehouses,
186
00:05:58,240 --> 00:05:59,760
the SQL Analytics engine.
187
00:05:59,760 --> 00:06:01,800
Now, if the lake house is the data engineering hub,
188
00:06:01,800 --> 00:06:04,120
think of the warehouse as the SQL Analytics engine,
189
00:06:04,120 --> 00:06:05,880
built for a different kind of user.
190
00:06:05,880 --> 00:06:07,920
A warehouse in fabric is a fully managed
191
00:06:07,920 --> 00:06:09,680
SQL Analytics database.
192
00:06:09,680 --> 00:06:12,240
If you're a database administrator or a SQL developer,
193
00:06:12,240 --> 00:06:13,840
this is where you'll feel at home.
194
00:06:13,840 --> 00:06:15,200
You can write T-Sycle queries.
195
00:06:15,200 --> 00:06:18,600
You can create views, stored procedures, and security policies.
196
00:06:18,600 --> 00:06:20,600
It looks and feels like a traditional data warehouse,
197
00:06:20,600 --> 00:06:22,040
but here's what makes it different.
198
00:06:22,040 --> 00:06:23,440
Unlike a traditional data warehouse,
199
00:06:23,440 --> 00:06:26,120
fabrics warehouse stores data directly in one lake.
200
00:06:26,120 --> 00:06:27,880
There's no separate storage layer,
201
00:06:27,880 --> 00:06:29,400
no extra infrastructure to manage.
202
00:06:29,400 --> 00:06:30,800
The data lives in one lake,
203
00:06:30,800 --> 00:06:33,640
and the warehouse is just the interface you use to query it.
204
00:06:33,640 --> 00:06:36,160
This matters because the warehouse and the lake house
205
00:06:36,160 --> 00:06:38,640
share the same data, so you're not duplicating anything.
206
00:06:38,640 --> 00:06:41,080
If a data engineer loads data into a lake house,
207
00:06:41,080 --> 00:06:43,760
a SQL developer can query it from a warehouse immediately,
208
00:06:43,760 --> 00:06:46,720
no copying, no moving, it's the same data accessed
209
00:06:46,720 --> 00:06:48,040
through different tools.
210
00:06:48,040 --> 00:06:51,000
The warehouse is built for large-scale analytics.
211
00:06:51,000 --> 00:06:53,200
Under the hood, it uses massive parallel processing
212
00:06:53,200 --> 00:06:54,960
to handle huge data sets.
213
00:06:54,960 --> 00:06:57,320
And Microsoft has been investing heavily in performance.
214
00:06:57,320 --> 00:06:58,920
In the last six months alone,
215
00:06:58,920 --> 00:07:01,520
performance has improved by 36%.
216
00:07:01,520 --> 00:07:03,200
That's a lot of improvement in a short time.
217
00:07:03,200 --> 00:07:04,920
If you're migrating from another platform,
218
00:07:04,920 --> 00:07:06,600
fabric has a migration assistant
219
00:07:06,600 --> 00:07:08,800
that helps you move from snowflake or synapse.
220
00:07:08,800 --> 00:07:10,320
It analyzes your existing setup
221
00:07:10,320 --> 00:07:11,800
and guides you through the process.
222
00:07:11,800 --> 00:07:14,200
The goal is to make migration as painless as possible.
223
00:07:14,200 --> 00:07:16,400
So what's the difference between a lake house and a warehouse?
224
00:07:16,400 --> 00:07:17,520
It comes down to the interface.
225
00:07:17,520 --> 00:07:19,680
We're houses are SQL first.
226
00:07:19,680 --> 00:07:21,640
They're built for people who think in terms of tables,
227
00:07:21,640 --> 00:07:23,480
queries and stored procedures.
228
00:07:23,480 --> 00:07:25,280
Lake houses are developer first.
229
00:07:25,280 --> 00:07:27,600
They're built for people who want to use Spark, Python,
230
00:07:27,600 --> 00:07:28,520
and notebooks.
231
00:07:28,520 --> 00:07:30,960
But underneath, they both store data in one lake.
232
00:07:30,960 --> 00:07:33,160
There are two ways of working with the same data.
233
00:07:33,160 --> 00:07:36,320
Getting data in, data factory, mirroring, and shortcuts.
234
00:07:36,320 --> 00:07:37,960
So you've got one lake for storage,
235
00:07:37,960 --> 00:07:40,720
lake houses for transformation and warehouses for querying.
236
00:07:40,720 --> 00:07:42,720
But how does the data actually get in?
237
00:07:42,720 --> 00:07:44,120
That's where data factory steps in.
238
00:07:44,120 --> 00:07:47,080
Data factory is Fabrics built in data integration service.
239
00:07:47,080 --> 00:07:48,560
And it gives you over 200 connectors
240
00:07:48,560 --> 00:07:51,880
to pull data from databases, cloud storage, SAS apps, and more.
241
00:07:51,880 --> 00:07:54,240
It's the most used data integration tool in the world
242
00:07:54,240 --> 00:07:55,400
and for good reason.
243
00:07:55,400 --> 00:07:57,920
Within data factory, you get data flows gen 2,
244
00:07:57,920 --> 00:08:00,040
which gives you a no-code way to transform data
245
00:08:00,040 --> 00:08:01,200
using power query.
246
00:08:01,200 --> 00:08:03,640
If you've ever used power query in Excel or Power BI,
247
00:08:03,640 --> 00:08:04,800
you already know the drill.
248
00:08:04,800 --> 00:08:06,240
You click through a visual interface
249
00:08:06,240 --> 00:08:08,840
to clean reshape and combine data, no coding required,
250
00:08:08,840 --> 00:08:11,520
and it's enterprise-grade, scalable to handle massive data
251
00:08:11,520 --> 00:08:12,240
sets.
252
00:08:12,240 --> 00:08:15,920
Over 22 billion orchestrations run on data factory every month,
253
00:08:15,920 --> 00:08:18,640
which means billions of data movements happen automatically,
254
00:08:18,640 --> 00:08:20,760
making it a serious piece of infrastructure.
255
00:08:20,760 --> 00:08:24,280
But there's an even simpler way to get data into one lake, mirroring.
256
00:08:24,280 --> 00:08:26,560
Mirroring copies data from external databases
257
00:08:26,560 --> 00:08:29,200
like Azure, School, Oracle, or Google BigQuery
258
00:08:29,200 --> 00:08:30,760
into one lake in near real time.
259
00:08:30,760 --> 00:08:32,280
And here's the best part.
260
00:08:32,280 --> 00:08:35,280
It's free, included in the service with no extra cost.
261
00:08:35,280 --> 00:08:36,800
You just point to your source database
262
00:08:36,800 --> 00:08:38,360
and Fabric handles the rest.
263
00:08:38,360 --> 00:08:40,560
Then there are shortcuts, which we touched on earlier,
264
00:08:40,560 --> 00:08:42,160
but they deserve more attention.
265
00:08:42,160 --> 00:08:44,840
Shortcuts are virtual pointers to data stored elsewhere,
266
00:08:44,840 --> 00:08:46,360
requiring no data movement.
267
00:08:46,360 --> 00:08:48,960
You create a shortcut to an AWS tree bucket,
268
00:08:48,960 --> 00:08:51,440
and it shows up in one lake as if it was stored there,
269
00:08:51,440 --> 00:08:52,920
while the data stays where it is.
270
00:08:52,920 --> 00:08:54,200
You're just pointing to it.
271
00:08:54,200 --> 00:08:56,880
And now shortcuts can do more than just point.
272
00:08:56,880 --> 00:08:59,840
Shortcut transforms let you apply AI-powered transformations
273
00:08:59,840 --> 00:09:00,680
on the fly.
274
00:09:00,680 --> 00:09:04,160
You can run sentiment analysis on text data, translate content,
275
00:09:04,160 --> 00:09:06,560
or detect personally identifiable information
276
00:09:06,560 --> 00:09:09,200
all without moving the data or building a pipeline.
277
00:09:09,200 --> 00:09:11,640
The transformation happens as you access it.
278
00:09:11,640 --> 00:09:13,600
The goal of all these tools is simple.
279
00:09:13,600 --> 00:09:16,360
Bring all your data into one lake without building complex
280
00:09:16,360 --> 00:09:19,520
pipelines, whether you use data factory for traditional ETL,
281
00:09:19,520 --> 00:09:21,480
mirroring for near real time replication
282
00:09:21,480 --> 00:09:24,840
or shortcuts for virtual access, the result is the same.
283
00:09:24,840 --> 00:09:29,120
Your data ends up in one lake ready for any Fabric workload.
284
00:09:29,120 --> 00:09:31,440
Real-time intelligence databases in AI.
285
00:09:31,440 --> 00:09:33,640
So far, we've talked about moving data and storing it,
286
00:09:33,640 --> 00:09:35,720
but what about data that never stops moving?
287
00:09:35,720 --> 00:09:38,640
Streaming data from IoT sensors, website click streams,
288
00:09:38,640 --> 00:09:41,520
telemetry from devices, that's where real-time intelligence
289
00:09:41,520 --> 00:09:42,240
comes in.
290
00:09:42,240 --> 00:09:44,440
Fabric handles streaming data through something called
291
00:09:44,440 --> 00:09:46,640
an event house, which stores and queries
292
00:09:46,640 --> 00:09:49,160
high volume time series data using KQL,
293
00:09:49,160 --> 00:09:50,720
the custochery language.
294
00:09:50,720 --> 00:09:53,000
It's the same engine that powers Azure Data Explorer
295
00:09:53,000 --> 00:09:56,000
built for speed, so you can ingest millions of events per second
296
00:09:56,000 --> 00:09:57,520
and query them in milliseconds.
297
00:09:57,520 --> 00:09:58,280
This matters.
298
00:09:58,280 --> 00:09:59,160
And here's why.
299
00:09:59,160 --> 00:10:02,520
46% of fabric customers already use real-time intelligence.
300
00:10:02,520 --> 00:10:03,600
That's nearly half.
301
00:10:03,600 --> 00:10:05,240
The use cases are everywhere.
302
00:10:05,240 --> 00:10:07,440
A logistics company tracking delivery trucks,
303
00:10:07,440 --> 00:10:09,480
a retailer monitoring in store for traffic,
304
00:10:09,480 --> 00:10:12,360
a manufacturer watching sensor data from factory equipment,
305
00:10:12,360 --> 00:10:14,200
all happening in real time.
306
00:10:14,200 --> 00:10:15,960
But fabric isn't just about analytics.
307
00:10:15,960 --> 00:10:17,680
It also includes operational databases
308
00:10:17,680 --> 00:10:19,440
like a SQL database for relational data
309
00:10:19,440 --> 00:10:21,440
and Cosmos DB for no-school workloads.
310
00:10:21,440 --> 00:10:24,520
These are the same databases you'd use to run your applications.
311
00:10:24,520 --> 00:10:26,120
And because they're built into Fabric,
312
00:10:26,120 --> 00:10:28,760
they automatically store their data in one lake.
313
00:10:28,760 --> 00:10:30,800
No extra pipelines, no manual copying,
314
00:10:30,800 --> 00:10:34,200
your transactional data and analytical data live in the same place.
315
00:10:34,200 --> 00:10:36,400
Now let's talk about AI because this is where fabric
316
00:10:36,400 --> 00:10:37,680
gets really interesting.
317
00:10:37,680 --> 00:10:40,000
AI integration runs deep across the platform
318
00:10:40,000 --> 00:10:42,120
and you get pre-built models for text analysis,
319
00:10:42,120 --> 00:10:43,880
translation and sentiment analysis.
320
00:10:43,880 --> 00:10:45,600
You can use them directly in your data flows
321
00:10:45,600 --> 00:10:48,160
without writing any code, just point to your data,
322
00:10:48,160 --> 00:10:49,680
select the model and it runs.
323
00:10:49,680 --> 00:10:52,440
And then there's Copilot, which is embedded across fabric.
324
00:10:52,440 --> 00:10:54,320
It helps you write queries, build reports,
325
00:10:54,320 --> 00:10:55,360
and create pipelines.
326
00:10:55,360 --> 00:10:57,400
You just describe what you want in plain English
327
00:10:57,400 --> 00:10:59,080
and Copilot generates the code.
328
00:10:59,080 --> 00:11:00,560
It's like having a junior data engineer
329
00:11:00,560 --> 00:11:02,960
sitting next to you ready to help at any moment.
330
00:11:02,960 --> 00:11:04,240
The vision here is simple.
331
00:11:04,240 --> 00:11:06,680
AI works on your data without moving it anywhere.
332
00:11:06,680 --> 00:11:09,640
Your data stays in one lake and the AI models come to the data.
333
00:11:09,640 --> 00:11:12,920
No copying, no exporting, no security risks.
334
00:11:12,920 --> 00:11:15,120
The data never leaves your governed environment.
335
00:11:15,120 --> 00:11:18,960
Power BI, governance and the business value.
336
00:11:18,960 --> 00:11:20,800
Now we get to the part most people actually see
337
00:11:20,800 --> 00:11:23,000
the front door to Insights Power BI.
338
00:11:23,000 --> 00:11:25,560
It's used by 550,000 organizations
339
00:11:25,560 --> 00:11:27,960
and 34 million users every month,
340
00:11:27,960 --> 00:11:30,680
making it the most widely used business intelligence tool
341
00:11:30,680 --> 00:11:31,640
in the world.
342
00:11:31,640 --> 00:11:34,120
And in fabric, it becomes something even more powerful.
343
00:11:34,120 --> 00:11:35,040
Here's the key.
344
00:11:35,040 --> 00:11:37,960
In fabric, Power BI connects directly to one lake
345
00:11:37,960 --> 00:11:39,920
using something called direct lake mode.
346
00:11:39,920 --> 00:11:41,640
This is different from the traditional import mode
347
00:11:41,640 --> 00:11:44,440
where you copy data into Power BI's internal storage.
348
00:11:44,440 --> 00:11:46,120
Instead, direct lake queries the data
349
00:11:46,120 --> 00:11:48,520
where it sits in one lake with no refresh needed,
350
00:11:48,520 --> 00:11:50,440
no duplication and the data is always current
351
00:11:50,440 --> 00:11:51,840
and blazing fast.
352
00:11:51,840 --> 00:11:55,120
Underneath every Power BI report is a semantic model.
353
00:11:55,120 --> 00:11:57,680
The brain of your report where you define business logic,
354
00:11:57,680 --> 00:11:59,360
measures and relationships.
355
00:11:59,360 --> 00:12:01,440
And now you can build these models entirely
356
00:12:01,440 --> 00:12:02,560
in a web browser.
357
00:12:02,560 --> 00:12:04,040
Web modeling is generally available
358
00:12:04,040 --> 00:12:06,720
so you can create a semantic model, build relationships
359
00:12:06,720 --> 00:12:09,560
and write DAX measures all from a browser even on a Mac.
360
00:12:09,560 --> 00:12:11,800
No desktop app required, but all this data power
361
00:12:11,800 --> 00:12:13,320
means nothing without governance.
362
00:12:13,320 --> 00:12:15,680
And fabric has governance built in from the start.
363
00:12:15,680 --> 00:12:17,960
Microsoft Perview provides sensitivity labels
364
00:12:17,960 --> 00:12:19,480
that follow your data everywhere.
365
00:12:19,480 --> 00:12:21,640
If a report contains sensitive information
366
00:12:21,640 --> 00:12:24,080
that labels stays with it, even when exported.
367
00:12:24,080 --> 00:12:26,680
Data loss prevention policies automatically detect
368
00:12:26,680 --> 00:12:29,800
sensitive data and block unauthorized sharing
369
00:12:29,800 --> 00:12:33,800
and access controls ensure only the right people see the right data.
370
00:12:33,800 --> 00:12:36,000
Domains help you organize data by business area
371
00:12:36,000 --> 00:12:37,480
with separate governance rules
372
00:12:37,480 --> 00:12:39,600
so your finance data can have stricter controls
373
00:12:39,600 --> 00:12:41,160
than your marketing data.
374
00:12:41,160 --> 00:12:43,120
Each domain can be managed by its own team,
375
00:12:43,120 --> 00:12:45,920
making it governance that scales with your organization.
376
00:12:45,920 --> 00:12:47,680
Now let's talk about real results.
377
00:12:47,680 --> 00:12:48,920
This isn't just theory.
378
00:12:48,920 --> 00:12:51,720
Eastman Chemical used fabric to reduce sales preparation
379
00:12:51,720 --> 00:12:54,960
time by 83%, what used to take four hours now
380
00:12:54,960 --> 00:12:56,400
takes about 40 minutes.
381
00:12:56,400 --> 00:12:59,240
Eaton Solutions cut manual effort by 75%
382
00:12:59,240 --> 00:13:02,520
with processes that require 10 to 15 steps now taking two.
383
00:13:02,520 --> 00:13:04,680
And Sonata Software saves 30,000 hours
384
00:13:04,680 --> 00:13:06,560
of manual reconciliation every year,
385
00:13:06,560 --> 00:13:09,240
redirecting that human effort to higher value work.
386
00:13:09,240 --> 00:13:10,960
The cost model is straightforward.
387
00:13:10,960 --> 00:13:14,200
You pay for capacity, the compute power that runs your workloads.
388
00:13:14,200 --> 00:13:16,400
Microsoft has built in smoothing and search protection
389
00:13:16,400 --> 00:13:17,480
to manage peaks.
390
00:13:17,480 --> 00:13:19,560
Smoothing spreads the cost of background operations
391
00:13:19,560 --> 00:13:22,480
over 24 hours, so you don't get hit with sudden spikes
392
00:13:22,480 --> 00:13:24,560
and search protection limits background jobs
393
00:13:24,560 --> 00:13:27,520
during busy periods to keep interactive performance stable.
394
00:13:27,520 --> 00:13:28,880
You pay for what you use
395
00:13:28,880 --> 00:13:31,080
and you don't get surprised.
396
00:13:31,080 --> 00:13:32,320
Let's bring this all together.
397
00:13:32,320 --> 00:13:34,640
Fabric isn't just another Microsoft product.
398
00:13:34,640 --> 00:13:37,520
It's a complete rethinking of how data platforms should work.
399
00:13:37,520 --> 00:13:41,040
One platform, one lake, one copy of data and infinite possibilities.
400
00:13:41,040 --> 00:13:43,760
And with AI and co-pilot built in, it's accessible to everyone,
401
00:13:43,760 --> 00:13:46,480
not just data professionals, but business analysts, sales teams,
402
00:13:46,480 --> 00:13:48,880
operations managers, anyone who needs answers from data.
403
00:13:48,880 --> 00:13:49,880
Here's your homework.
404
00:13:49,880 --> 00:13:52,960
Go to fabric, Microsoft.com, start a free trial
405
00:13:52,960 --> 00:13:54,480
and create your first workspace.
406
00:13:54,480 --> 00:13:55,280
Just try it.
407
00:13:55,280 --> 00:13:57,440
You'll see how the pieces fit together in a way
408
00:13:57,440 --> 00:13:59,440
that no amount of reading can replace.
409
00:13:59,440 --> 00:14:01,560
If this episode helped you understand fabric better,
410
00:14:01,560 --> 00:14:03,320
subscribe on your favorite podcast platform
411
00:14:03,320 --> 00:14:05,600
and share it with someone who's starting their data journey.
412
00:14:05,600 --> 00:14:06,520
They'll thank you.
413
00:14:06,520 --> 00:14:10,320
Next time, we'll break down one lake short cuts in plain English.
414
00:14:10,320 --> 00:14:12,200
How they work, why they matter,
415
00:14:12,200 --> 00:14:14,840
and how to use them without moving data around.
416
00:14:14,840 --> 00:14:16,080
That's one you won't want to miss.
417
00:14:16,080 --> 00:14:17,480
Until then, keep learning.
