Ad
More Related Content
What's hot (20)
Similar to Azure Analysis Services (Azure Bootcamp 2018) (20)
Ad
Recently uploaded (20)
Ad
Azure Analysis Services (Azure Bootcamp 2018)
- 2. Introduction • Education • B.S. in Computer Science from UW-Milwaukee • M.S. in Science of Management from Cardinal Stritch • Professional • Fiserv, DentaQuest, BMO Harris • Focus on SQL, SSIS, and DevOps strategies • Talavant • Focus on larger BI strategies Turner Kunkel [email protected] Senior Consultant @ Talavant • Hobbies • Home brewing • Bass playing (poorly) • Sailing (light skipper license)
- 3. There is a better way to make data work for companies. Better resources, strategy, sustainability, inclusion of the organization as a whole, understanding of client needs, tools, outcomes, better ROI. • Accelerated planning, implementation and results • Sustainable solutions • Increased company-wide buy-in & usage VALUE WE PROVIDE By providing a holistic approach inclusive of a client’s people, processes and technologies - built on investment in our own employees and company growth. HOW WE DO IT STRATEGY ARCHITECTURE IMPLEMENTATION
- 4. OVERVIEW On-Premises SSAS History of AS • ROI considerations • Future research • Questions/Comments • Sources & References Azure AS Developing w/ AAS Automation on AAS
- 5. END PRODUCT Robust Power BI sample from Microsoft 2.04 billion NY Taxi trips Refresh using 20 Azure cores in ~4 seconds
- 7. OLAP Database theory (X,Y,Z) → W (X,Y,Z are axis in a cube, W is value of cell) Need arose from tab report structure from 1980’s database management systems Slice, Dice, Drill Down, Roll Up
- 9. MODELS FUNCTIONS Multidimensional Tabular Actions Yes No Aggregations Yes No Calculated Column No Yes Calculated Measures Yes Yes Calculated Tables No Yes 1 Custom Assemblies Yes No Custom Rollups Yes No Default Member Yes No Display folders Yes Yes 1 Distinct Count Yes Yes (via DAX) Drillthrough Yes Yes (depends on client application) Hierarchies Yes Yes KPIs Yes Yes Linked objects Yes Yes (linked tables) M expressions No Yes 1 Many-to-many relationships Yes No (but there is bi-directional cross filters at 1200 and higher compatibility levels) Named sets Yes No Ragged Hierarchies Yes Yes 1 Parent-child Hierarchies Yes Yes (via DAX) Partitions Yes Yes Perspectives Yes Yes Row-level Security Yes Yes Object-level Security Yes Yes 1 Semi-additive Measures Yes Yes Translations Yes Yes User-defined Hierarchies Yes Yes Writeback Yes No Multidimensional SSAS Power BI (in preview) Tabular SSAS, Azure Power BI, Power Pivot, etc. Created from Multidimensional metadata TMSL code TOM code Compatibilities: 1050, 1103, 1200, 1400 Azure >= 1200 level
- 10. ARCHITECTURE
- 11. BUSINESS USE • Explore data from outside the organization • Historical insight • Trends and predictive analysis • Self-service opportunities for business users • Space saving • Time saving Reporting Tools SSRS SharePoint Power Pivot Power BI Tableau (and other third parties)
- 13. SETUP ON-PREMISES Infrastructure Setup time, Developers, Space, Age, Networking Administration Security, Scaling, Disaster recovery, Licensing, Local backups
- 15. SETUP Initial thoughts Where in Azure is it? How do I start it? What is the cost?
- 16. INSTANCE CREATION Resource Group Query Processing Units (QPU) Location Backup Storage
- 17. INSTANCE UPKEEP Deployment Scripts Instance metrics Security Troubleshooting Scaling Backups
- 19. MODEL DEVELOPMENT Web interface (public preview) Edit relationships, measures, hierarchies Drag-and-drop query editor for data Translates to DAX if needed Interface to open model in several tools SQL Server Data Tools/Visual Studio (>=2016) SQL Server Management Studio (>=2016)
- 20. MIGRATION Best practice Incremental port of solution Lean, Phased, Ramp up Types Full Hybrid/piece-wise Tools On-premises Gateway Visual Studio, SQL Management Studio
- 21. GATEWAY Install Download, install, register on-premises Use Connect gateway to AAS instance Configure Add gateway resource to Azure *Performs slowly on wireless *Communication is encrypted
- 22. DATA ACCESS DirectQuery benefits Up-to-date data Stronger security Optimized query plan In-Memory Cache Stronger query performance Data refresh required DirectQuery limitations Sources SQL server, Oracle, Teradata No stored procedures No calculated tables Query language translations
- 23. “HIGH AVAILABILITY” Backups Azure allows backup storage Configurable in the AAS instance Can backup and restore to separate instances Hint: Use deployment scripts Redundancy Rarely, Azure servers go down Ensure availability by deploying to another instance in a different region Process each instance in parallel
- 25. AUTOMATION Automation Uses • Model processing on schedule • Full or incremental • Start/Stop AAS on schedule • Scale AAS automatically • Backup AAS instance on schedule How to Automate • Azure Function Apps • Azure RunBooks • Classic SSMS/SSIS • Custom .NET application • Custom PowerShell
- 26. FUNCTION APPS • Separate module in Azure • Premade Functions • Webhook/API, Timer (CRON), Data Processing • Templates (C#, F#, JavaScript) • Custom functions • PowerShell, Python, and Windows Batch • Analysis Server libraries, providers, and connection string references needed • Web application sitting in Azure • Separate pricing model • Can use Source Control/TFS
- 27. RUNBOOKS • Requires Azure Automation account • Uses ‘Run As’ account to connect to AAS • RunBooks support PowerShell and Python scripts • Can be run on a recurring schedule
- 29. ROI & CONSIDERATIONS Advantages to movement to AAS • No physical space needed for servers • One cloud platform for development • Azure is Microsoft’s future – AAS is beneficial to learn • Scaling and automating use is quick • Integration to already used tools • Processing power is immense for large data sets • Access administration is streamlined • Redundancy built-in • Simple and forecasted pricing model Disadvantages to movement to AAS • Learning curve • (Possible!) replacement of SSAS resume skill
- 30. OTHER FEATURES Query Replicas Queries distributed among multiple replicas in a pool Up to 7 additional query streams (8 total) Good for high-QPU usage times Can be configured based on instance usage Tags Azure general feature – Name/Value pair for resources For example: Environment/Development Locks Azure general feature – Locks resource to prevent actions on the resource Diagnostics Performance logs Event logs Error logs Azure infrastructure logs
- 31. FUTURE FEATURES AAS support in Power BI Embedded Azure Data Lake storage as data source Multidimensional cube support Schema compare Excel Online connections AAS data source in SSRS
- 32. FUTURE RESEARCH Automation best practices Deterministic scaling development Best uses of RunBooks, Functions, Replicas, or other techniques? Migration strategy Full approach to migration plan Testing techniques while migrating leanly True Implementation Cost Production down time while migrating Opportunity cost and monetary cost during learning curve Buy-in from business On-boarding of business users, if necessary Production monitoring Best practices for maintenance How many resources required Improvements to implemented system and development life cycle
- 33. RECAP On-Premises SSAS History of AS • ROI considerations • Future research • Questions/Comments • Sources & References Azure AS Developing w/ AAS Automation on AAS
- 34. QUESTIONS?
- 35. RESOURCES • AAS product page: https://azure.microsoft.com/en-us/services/analysis-services/ • MS Docs overview: https://docs.microsoft.com/en-us/azure/analysis-services/analysis-services-overview • AAS blog: https://blogs.msdn.microsoft.com/analysisservices/ • AAS Web Designer: https://docs.microsoft.com/en-us/azure/analysis-services/analysis-services-create-model- portal • Feature requests: https://feedback.azure.com/forums/556165?status_id=191761 • Third party/community code: http://sqlsrvanalysissrvcs.codeplex.com/ • Video overviews: https://channel9.msdn.com/series/Azure-Analysis-Services • Video introduction (using NY Taxi data): https://azure.microsoft.com/en-us/resources/videos/introducing-azure- analysis-services/ • NY Taxi data (lots!): https://chriswhong.com/open-data/foil_nyc_taxi/