Can you Re-Platform your Teradata, Oracle, Netezza and SQL Server Analytic Workloads to Hadoop?
4 likes1,005 views
The document discusses re-platforming existing enterprise business intelligence and analytic workloads from platforms like Oracle, Teradata, SAP and IBM to the Hadoop platform. It notes that many existing analytic workloads are struggling with increasing data volumes and are too costly. Hadoop offers a modern distributed platform that can address these issues through the use of a production-grade SQL database like VectorH on Hadoop. The document provides guidelines for re-platforming workloads and notes potential benefits such as improved performance, reduced costs and leveraging the Hadoop ecosystem.
Can you Re-Platform your Teradata, Oracle, Netezza and SQL Server Analytic Workloads to Hadoop?
- 1. Confidential © 2014 Actian Corporation1 Confidential © 2015 Actian Corporation1 Confidential © 2015 Actian Corporation Re-platforming Enterprise BI & Analytic Workloads to Hadoop Mike Hoskins, CTO June 2015 AKA: Can I move analytic workloads off Oracle, Teradata, SAP, IBM and Microsoft to Hadoop??
- 2. Confidential © 2014 Actian Corporation2 Confidential © 2015 Actian Corporation2 Agenda • Who is Actian • Key Trends Shaping Big Data • Existing Analytic Workloads are “Hitting the Wall” • Can I Re-platform to Hadoop? • After Re-platforming, then What? • Summary
- 3. Confidential © 2014 Actian Corporation3 Confidential © 2015 Actian Corporation3 Who is Actian? $140M Revenues + Profitable 10,000+ Customers Global Presence: 8 world-wide offices, 7x 24 multinational support model End to End Big Data Platform. Disruptive Price Performance. 3 “Actian Analytics demonstrates that the company now has an impressive range of offerings that have been rebranded and combined in a pretty comprehensive framework.” 451 Research “Actian is now very powerfully positioned in the big data and data analytics markets.” Bloor Group 3
- 4. Confidential © 2014 Actian Corporation4 Confidential © 2015 Actian Corporation4 SQL BI & Analytics Actian AAP Supporting: Kafka, HBase, Ambari, YARN, Spark, Ranger, Atlas “Vortex” on Hadoop Predictive Analytics The Wiz Data Scientist IT Sophisticate CIO Maestro Business Analyst Speed Demon Actian Vortex: Analytic Pipelines for Big Data
- 5. Confidential © 2014 Actian Corporation5 Confidential © 2015 Actian Corporation5 Key Trends • Modern Software • Fully distributed data & compute, on commodity HW • Elastic, Super-scaling, Hyper-parallelized • The Rise of Analytic Workloads • Hadoop is pushing the Analytic Frontier • But what about existing Enterprise BI & Analytic Workloads?
- 6. Confidential © 2014 Actian Corporation6 Confidential © 2015 Actian Corporation6 Re-platforming BI & Analytic Workloads to Hadoop ISVs Custom Apps
- 7. Confidential © 2014 Actian Corporation7 Confidential © 2015 Actian Corporation7 Enterprise Analytic Workloads • Enterprise Analytic Workloads are pervasive • They are gaining Strategic Value, since Business Value increasingly comes from this “analytic tier” • Grow revenue via better marketing and customer analytics • Reduce cost via better fraud detection and payment analytics • Mitigate risk via better compliance and cyber analytics
- 8. Confidential © 2014 Actian Corporation8 Confidential © 2015 Actian Corporation8
- 9. Confidential © 2014 Actian Corporation9 Confidential © 2015 Actian Corporation9 But Enterprise Analytic Workloads are Starting to “Break” • We hear from our Customers and Prospects that many of these Enterprise Analytic Workloads are “breaking” • Performance is not keeping up with data volumes • Too slow for modern iterative analytic techniques • Too distant from the new data sources that are driving new analytic insights • Too costly • Built on 35 year-old technology and architecture • These users often feel “trapped” in a dead-end, what is the cause?
- 10. Confidential © 2014 Actian Corporation10 Confidential © 2015 Actian Corporation10 In Enterprise Analytic Workloads Databases are the Weak Link • Big Data is exposing the legacy “database tier” as the weak link in existing Enterprise Analytic Workloads • Can’t deal with volume • Can’t deliver on the promise of new advanced analytics • Bottom line: legacy Databases don’t perform for modern analytic workloads • Why are they failing? • Why is this failure so punishing to their customers? • The Performance “Edge” is critical to analytic success
- 11. Confidential © 2014 Actian Corporation11 Confidential © 2015 Actian Corporation11 Enterprise Analytic Workloads – What to do? • How to preserve (and even grow) the huge Value contained in this ubiquitous “analytic tier”? • How to optimize and future-proof these vital Enterprise Analytic Workloads? • Is there a “modern” platform out there to save the day?
- 12. Confidential © 2014 Actian Corporation12 Confidential © 2015 Actian Corporation12 Re-platforming BI & Analytic Workloads to Hadoop Legacy HW & SW Platforms ISVs Custom Apps QL
- 13. Confidential © 2014 Actian Corporation13 Confidential © 2015 Actian Corporation13 Can my Enterprise BI & Analytics be Re-platformed to Hadoop? • Yes! • The key is SQL • Thousands of SQL Tools and Apps, and millions of vital SQL workloads are currently plugged into legacy databases that are losing the crucial price/performance battle • Many of these SQL analytic workloads are candidates to “re-platform” to a modern SQL database running in Hadoop
- 14. Confidential © 2014 Actian Corporation14 Confidential © 2015 Actian Corporation14 But is Hadoop Ready for Enterprise-class SQL Workloads? • Hadoop is still very immature when it comes to SQL • The missing piece is a production-grade, enterprise class, columnar analytic database • Surveying the SQL-in-Hadoop landscape: • Legacy players? • Existing SQL-in-Hadoop offerings? • Wrapped Legacy • Immature or Query-only
- 15. Confidential © 2014 Actian Corporation15 Confidential © 2015 Actian Corporation15 Enter VectorH: Enterprise-class, Low-latency SQL-in-Hadoop • Established and Open • Deep innovation yields extreme high-performance • Inventors of vector-processing in modern databases • Persistent, ACID and full CRUD • Enterprise capabilities • Full SQL compliance • Plug-n-play your favorite SQL Client (it just works! And lightning fast)
- 16. Confidential © 2014 Actian Corporation16 Confidential © 2015 Actian Corporation16 Re-platforming BI & Analytic Workloads to Hadoop Legacy HW & SW Platforms ?Actian Vortex w/ VectorH ISVs Custom Apps QL
- 17. Confidential © 2014 Actian Corporation17 Confidential © 2015 Actian Corporation17 Looking at some Re-platforming Cases: • Top global retail bank • Re-platform from Oracle to VectorH • Top media company • Re-platform from Teradata to VectorH • Top healthcare company • Re-platform from SQL Server to VectorH • And others..
- 18. Confidential © 2014 Actian Corporation18 Confidential © 2015 Actian Corporation18 Re-Platforming Guidelines • Avoid “rip and replace” (we are not replacing the legacy database, we are optimizing and future-proofing key analytic workloads by moving them to a modern database on a modern platform) • Watch where Business Logic resides? • Set up data interfaces • One-time data migration • Ongoing one-way replication? • Leverage existing or new database load pipelines?
- 19. Confidential © 2014 Actian Corporation19 Confidential © 2015 Actian Corporation19 Re-Platforming Guidelines (Part 2) • Target the most “broken” analytic workloads • Analytic apps? • Enterprise reporting? • Ad-hoc query/BI and data discovery? • DM/DW/EDW? • And ideally those with the “cleanest” SQL
- 20. Confidential © 2014 Actian Corporation20 Confidential © 2015 Actian Corporation20 Tailwinds from Re-platforming to Vector on Hadoop • Moving compute to where the data lives • Fully open data tier (the end of vendor lock-in!) • Benefit from deep innovation in Ingest Frameworks • Better equipped to handle streaming/realtime sources • Ability to exploit more unstructured data sources
- 21. Confidential © 2014 Actian Corporation21 Confidential © 2015 Actian Corporation21 Tailwinds from Re-platforming to Vector on Hadoop (Part 2) • Better support for the imminent IoT explosion • Leverage the burgeoning Apache ecosystem • Make world-class SQL available to ALL your existing and future Hadoop workloads • Tap into a rich vein of Hadoop “veterans” who can enable and even accelerate the next great wave of Hadoop adoption (*yes, that’s you folks in the room!)
- 22. Confidential © 2014 Actian Corporation22 Confidential © 2015 Actian Corporation22 Re-platforming BI & Analytic Workloads to Hadoop Legacy HW & SW Platforms ?Actian Vortex w/ VectorH ISVs Custom Apps QL
- 23. Confidential © 2014 Actian Corporation23 Confidential © 2015 Actian Corporation23 Summary • The Early Majority wave is about to sweep over Hadoop, and one of the key drivers will be the re-platforming of existing SQL-based Enterprise Analytic Workloads to Hadoop • The key enabler is richness and performance of your SQL-in-Hadoop • Make the right SQL-in-Hadoop choice and speed your transition from the last 30 years to the next 30 years!
- 24. Confidential © 2014 Actian Corporation24 Confidential © 2015 Actian Corporation24 www.actian.com facebook.com/actiancorp @actiancorp Thank You
Editor's Notes
- #21: These goodies aren’t why you make the move. Do so because the workloads are important, but broken. You have to jump. You are in a burning building. Good news is, once you are there, swimming in Hadoop swimming pool, you’ve jumped into a future filled with tailwinds. No IoT will be written for Teradata.
- #22: These goodies aren’t why you make the move. Do so because the workloads are important, but broken. You have to jump. You are in a burning building. Good news is, once you are there, swimming in Hadoop swimming pool, you’ve jumped into a future filled with tailwinds. No IoT will be written for Teradata.