--session donnée lors du SQL Saturday Torino 2015--
http://www.sqlsaturday.com/454/EventHome.aspx
Data Management Gateway (classic, personal, AS connector) is what make modern Microsoft BI stack hybrid. Power BI (v1 and v2) and Azure Data Factory use that component to interact with On-Prem Data assets.
That session is a Deep dive into the DMG and the hybrid architecture involved by Power BI and ADF. How does it work ? Security, Firewall, Certificates, Multiple gateways, Admin delegation, Scale out, Disaster Recovery…. All that topics will be covered during that technical session.
3. #SQLSAT454
Jean-Pierre Riehl
Practice Manager Data & BI
@djeepy1 - http://blog.djeepy1.net
Who am I ?
Pure-Player Microsoft
Self-Service BI experts
« Agile BI » projects
Data Strategy
Hybrid-BI
MVP SQL Server
French PASS
Chapter Leader
4. #SQLSAT454
La communauté Data Microsoft
Webcasts, Conférences, Afterworks
.Pro
Save the Date
11/30th & 12/1st
Campus Microsoft
Paris
Self-promotion
2 days
40 sessions
50 speakers
100% Free
5th edition
10. #SQLSAT454
One DMG to rule them all ?
DMZ
SQL
Cloud On-Prem
Oracle
…
Power BI for O365
Data
Management
Gateway
OnPrem Data Sources
PowerBI.com
Azure Data Factory
Excel
11. #SQLSAT454
Versions of DMG
There are multiple versions of DMG
Data Management Gateway
Power BI (v1)
ADF
1.6.5735.1 23/9/2015
Power BI Personal Gateway Power BI 1.6.5715.1 3/9/2015
Analysis Services Connector Power BI 1.6.5696.2 2/9/2015
All with the same base code
12. #SQLSAT454
Power BI v1
(for O365)
31/12
24/07
Power BI v2
powerbi.com
Azure
Data Factory
Classic DMG ??
SSAS Connector
Personal Gateway
Versions timeline
Classic DMG ??
14. #SQLSAT454
Power BI
Excel
Power Query Power Pivot
Power View Power Map
Power BI Tenant SQL
Cloud On-Prem
Oracle …
Power BI
Admin Center
Data
Catalog
Q&A
Power BI Sites
Data Refresh
Indexing / Search
OData
Access
19. #SQLSAT454
What’s that
A Windows Service
DIAHostService (Data Management Gateway)
A configuration application
Managed by the cloud
Power BI Admin Center
Azure Portal (new version)
Manage by User
An application
26. #SQLSAT454
Data Source Discovery
Data Sources are declared on « Admin
Center »
They are associated to a Gateway
Power BI Service resolves Data Source and
find right Gateway
Resolution = Provider + Server + Catalog
Warning: Case sensitive
Personal
& Self-Registered
27. #SQLSAT454
Cloud On-Prem
Data Refresh explained
Power BI
OnPrem
Sources
DMG
Azure Service
Bus
0
Register DMG
Register Data Sources
1
Ask for Data Refresh
Check permissions
Find Gateway
4 Execute Query
5 Get Data
6 Compress, Chunk
Queue
request2
3 Read
request
Gateway
Cloud Service
7 Update Model
28. #SQLSAT454
Cloud On-Prem
OData access explained
Power BI
OnPrem
Sources
DMG
0 Index Sources
2
Resolve Data Source
Redirect to local Gateway
4 Execute Query
3 Request Data
1 Search Source
Excel
Power Query
Hybrid Proxy
30. #SQLSAT454
Excel
Power Query
Power BI
DMG
Data sources
• SQL (1433)
• FileShare
• Etc.
+DNS resolution
Azure Service Bus (outbound)
• 9350-9354 / 5671,5672
Power BI Admin Center (outbound)
• 80/443
Direct HTTP
(inbound)
• HTTPS (8050)
• HTTP (8051)
HTTPS
(443)
Cloud
On-Prem
Network Security
Credential
Manager
32. #SQLSAT454
Impersonation by DMG
Connection identity is impersonated by DMG
Peer-to-peer connection
Credentials are stored by DMG for each
source
Credentials can be stored encrypted locally
Credentials can be stored encrypted on Cloud
Credentials are set up locally
Credential Manager is a Click Once application
Credential Manager discuss with DMG
Or set up on the cloud
34. #SQLSAT454
Impersonation by the Source
Used by Live Query
Keep identity of original user (Kerberos-like)
Use the EffectiveUsername method
DMG authenticates to SSAS
DMG passes EffectiveUsername property
SSAS executes the query as if it was original user
DMG must be SSAS Admin
UPN must be declared in AD
37. #SQLSAT454
As easy as scale-out
DMGs
Cloud
On-Prem
Power BI
…
Instance 1 Instance 2
Server A Server B
38. #SQLSAT454
HA & DR
Gateways and Data sources are managed by
Cloud
Credentials are encrypted by MDG
Use of Certificate, self-generated or yours
Need Certificate when registering a new instance
Warning when stored localy
Load Balancing
40. #SQLSAT454
Last thought
« Classic » DMG
IT managed
Data source granularity
Data sources registration
Credentials bound to source
HA & DR
Personal Gateway
User Managed
User Granularity
Self data sources registration
Credentials bound to user
No service continuity
Waiting for merge !!!!!!
42. #SQLSAT454
Resources
Getting Started : https://support.office.com/en-us/article/Power-BI-for-Office-365-Admin-Center-Help-
5e391ecb-500c-47a3-bd0f-a6173b541044
Data Factory Reference : https://msdn.microsoft.com/en-us/library/azure/dn834987.aspx
Supported Data sources : https://support.office.com/en-us/article/Supported-Data-Sources-and-Data-
Types
DMG Prerequisites : https://support.office.com/en-us/article/Data-source-prerequisites-Power-Query
SSIS Data Feed : http://www.microsoft.com/en-us/download/details.aspx?id=39931
How does DMG work - sequence diagram :
https://support.office.com/en-us/article/An-Inside-look-at-Microsoft-Data-Management-Gateway-
078D306D-2D28-409C-962B-D80B69A1D0B2
Azure IP adresses : http://www.microsoft.com/en-us/download/details.aspx?id=41653
Store encrypted credentials in the cloud for business continuity : https://support.office.com/en-
us/article/Business-Continuity-for-a-Data-Management-Gateway-00106c58-a0a1-4cea-8886-
da51b8cadc5b
Matt Masson Session at SPC : http://channel9.msdn.com/Events/SharePoint-
Conference/2014/SPC397
Mini Nay & Luming Han Session at PASS Summit :
http://www.sqlpass.org/summit/2014/Sessions.aspx#sid7231
-Cloud First Strategy, just listen to Satya
-Innovation is in the cloud
-Examples : PBI, AzureML, HDInsigth, etc.
-If you want new technologies you probably will have to use a cloud-based one
-But…
-you can put all your data assets in the cloud, even,
-that will take a long time
-or it won’t never happen (mindset, investment, legal constraints, privacy…)
-So what
-Microsoft knows about that
-They build hybrid cloud, hybrid solutions
-They make investments in a technology to remove barriers
-Data Management Gateway
-it is the link between Cloud technologies and On-Prem Data
ADF & Power BI merged in march
-first, data refresh is trivial. Models are in the cloud, Data is on-prem.
-second, indexing data sources into data catalog
-third, relating Data Catalog, agnostic access to Data
-
Initiated by Cloud
Otherwise, there is no challenge, you can push data from On-Prem (SSIS, Power BI API, etc.)
Data sources :
Broken -> https://support.office.com/en-us/article/Supported-Data-Sources-and-Data-Types
French -> https://support.office.com/fr-FR/article/Sources-de-donn%C3%A9es-et-types-de-donn%C3%A9es-pris-en-charge-CB69A30A-2225-451F-A9D0-59D24419782E
Prerequisites : https://support.office.com/en-us/article/Data-source-prerequisites-Power-Query
SSIS Data Feed : http://www.microsoft.com/en-us/download/details.aspx?id=39931
Personal Gateway : https://support.powerbi.com/knowledgebase/articles/649846
Personal Data Sources : https://support.powerbi.com/knowledgebase/articles/474669-refresh-data-in-power-bi