Five ways database modernization simplifies your data life
Download as PPTX, PDF1 like2,470 views
This document provides an overview of how database modernization with MemSQL can simplify a company's data life. It discusses five common customer scenarios where database limitations are impacting data-driven initiatives: 1) Slow event to insight delays, 2) High concurrency causing "wait in line" analytics, 3) Costly performance requiring specialized hardware, 4) Slow queries limiting big data analytics, and 5) Deployment inflexibility restricting multi-cloud usage. For each scenario, it provides an example customer situation and solution using MemSQL, highlighting benefits like real-time insights, scalable user access, cost efficiency, accelerated big data analytics, and deployment flexibility. The document also introduces MemSQL capabilities for fast data ingestion, instant
More Related Content
Similar to Five ways database modernization simplifies your data life (20)
Five ways database modernization simplifies your data life
- 1. Five Ways Database Modernization Simplifies Your Data Life 1 Mike Boyarski, Senior Director Product Marketing
- 2. 2 ● Quick Database Landscape Roundup ● Five (Customer) Ways to a Simpler Data Life ● An Intro to MemSQL Agenda
- 3. 3 “Companies with data-driven environments have up to 50% higher market value than other businesses.”
- 4. 4 Data demands More people want data access, impacting database concurrency and scale Analytic performance Faster, more frequent intra-day insights Rising data management costs Costly upgrades and maintenance with existing technology Database Limitations are Impacting Data- Driven Initiatives
- 7. ● 15 Different Databases or DB Options ● Some are features ● Some are platforms
- 9. 9 Workload Database Type Transactions/Writes? Operational Queries/Reads? Analytical Both transactions and queries? HTAP (next slide)
- 10. An Intro to HTAP Hybrid Transaction/Analytical Processing “HTAP architectures remove the latency between when a piece of data is generated and when it is ready for analytics” 10 “Delivering Digital Business Value Using Practical Hybrid Transactional/Analytical Processing”, Analysts: Adam M. Ronthal, Roxane Edjlali
- 11. An Intro to HTAP Hybrid Transaction/Analytical Processing “HTAP architectures remove the latency between when a piece of data is generated and when it is ready for analytics” “The traditional batch-oriented cycle of business intelligence insight is giving way to new deployment architectures delivering real-time access to data and the expected insights that it will provide.” 11 “Delivering Digital Business Value Using Practical Hybrid Transactional/Analytical Processing”, Analysts: Adam M. Ronthal, Roxane Edjlali
- 12. An Intro to HTAP Hybrid Transaction/Analytical Processing “HTAP architectures remove the latency between when a piece of data is generated and when it is ready for analytics” “The traditional batch-oriented cycle of business intelligence insight is giving way to new deployment architectures delivering real-time access to data and the expected insights that it will provide.” “To support this demand, the convergence of analytic and operational platforms is becoming more common. “ 12 “Delivering Digital Business Value Using Practical Hybrid Transactional/Analytical Processing”, Analysts: Adam M. Ronthal, Roxane Edjlali
- 13. 13 Workload Database Type Transactions? Operational Queries? Analytical Both transactions and queries? HTAP (will explain next slide) Lots of transactions per second? 10,000+ -> Millions In-Memory Boosted Operational Fast reads + ad-hoc? Milliseconds on 100Ms/Bs of rows In-Memory boosted +Columnstore Analytical Data Science/ML Analytical or Data Lake
- 14. 14 Five Customer Scenarios Faster Event to Insight Growth in Concurrency Cost Effective Performance Accelerate Big Data Deployment Flexibility
- 15. 15 Event to Insight Delays InsightEvent Hours to Days
- 16. 16 Event to Insight Delays Common Challenges ● Protect performance limiting ad-hoc aggregates ● Load data during “non-peak” operations ● Writes slow down reads
- 17. 17 Faster Event to Insight Example Situation: Oracle Financials database could only support once/day reporting resulting in fraudulent and duplicate orders Industry: Energy
- 18. 18 Faster Event to Insight Example Solution: Real-time synchronization from Oracle to MemSQL with change data capture enabling continuous reporting Result: Cost savings Situation: Oracle Financials database could only support once/day reporting resulting in fraudulent and duplicate orders Industry: Energy
- 21. 21 DB Modernization for Faster Event to Insight ● Lock-free ingestion ● Distributed for faster parallel loading ● Scalable durability for reliability and accuracy ● Relational SQL for BI tool compatibility
- 22. 22 Five Customer Scenarios Faster Event to Insight Growth in Concurrency Cost Effective Performance Accelerate Big Data Deployment Flexibility
- 23. 23 “Wait in Line” Analytics Common Challenges ● Under-powered database can’t support data and user growth ● Too many non-standard ad-hoc queries ● Transactions impact queries
- 24. 24 High Query Concurrency Example Situation: Operational dashboards couldn’t support demand of roughly 250k queries across 1,000+ users Customer: UBER
- 25. 25 High Query Concurrency Example Solution: MemSQL delivers sub- second query response for 1,000s of users while supporting continuous writes Result: Real-Time Dashboards Situation: Operational dashboards couldn’t support demand of roughly 250k queries across 1,000+ users Customer: UBER
- 26. 26 Architecture Source: UBER Engineering Video at memsql.com/uber
- 27. 27 DB Modernization for Concurrency Growth ● Distributed architecture simplifies scale demands ● Query compilation for fast continuous queries ● In-memory optimized tables for continuous ingestion
- 28. 28 Five Customer Scenarios Faster Event to Insight Growth in Concurrency Cost Effective Performance Accelerate Big Data Deployment Flexibility
- 29. 29 Costly Performance Common Challenges ● Specialized HW required for growing data ● Costly DB options and accelerators ● Professional service tuning
- 30. 30 Costly Performance Example Situation: Degrading query and ingestion performance couldn’t deliver daily billing requirement resulting in lost revenue Industry: High Tech
- 31. 31 Costly Performance Example Situation: Degrading query and ingestion performance couldn’t deliver daily billing requirement resulting in lost revenue Industry: High Tech Solution: Accelerated ingestion from 56k/s to 6M/s 100x faster queries ⅓ the cost of incumbent solution Result: Added revenue
- 34. 34 DB Modernization for Cost Efficient Performance ● Optimized for industry standard hardware ○ Query vectorization - CPU acceleration ○ Scale-out architecture ● Columnar data compression - 5-10x ● Memory optimization at no additional license cost
- 35. 35 Five Customer Scenarios Faster Event to Insight Growth in Concurrency Cost Effective Performance Accelerated Big Data Deployment Flexibility
- 36. 36 Accelerate Big Data Common Challenges ● Slow queries ● Existing BI tools not compatible ● Flexible data structure hard to understand/use
- 37. 37 Accelerate Big Data Example Situation: Queries taking hours to days to complete limiting use of analytics for decisions Customer: Pandora
- 38. 38 Accelerate Big Data Example Solution: MemSQL delivers sub- second queries on 100s of billions of rows Result: Frequent use of analytics Situation: Queries taking hours to days to complete limiting use of analytics for decisions Customer: Pandora
- 40. 40 DB Modernization for Faster Big Data ● Lock-free architecture for fast ingestion ● Disk-based compute for cost effective storage ● Connectivity for bulk and stream loading (ie. Kafka, Spark, HDFS, S3, Azure Blob) ● Columnar table format eliminates pre- aggregations and background queries
- 41. 41 Five Customer Scenarios Faster Event to Insight Growth in Concurrency Cost Effective Performance Accelerated Big Data Deployment Flexibility
- 42. 42 Deployment Inflexibility On-Premises Cloud Hybrid Common Challenges ● Database runs on one cloud ● Cloud and On- Premises product versions are different
- 43. 43 Deployment Flexibility Example Situation: Cloud application deployed on Azure and Equinix for customer security requirements, requiring multi-cloud database Customer: Kollective
- 44. 44 Deployment Flexibility Example Situation: Cloud application deployed on Azure and Equinix for customer security requirements, requiring multi-cloud database Customer: Kollective Solution: MemSQL used on Azure and Equinix, achieving breakthrough performance for real- time dashboards Result: Improved customer experience
- 46. 46 The Modern Database for Today’s Performance Low Latency Queries Scalable ANSI SQL Petabyte scale Columnar compression Scalable User Access Scale-out for performance Converged transactions and analytics Multi-threaded processing Fast Loading Stream data On-the-fly transformation Multiple sources
- 47. MemSQL: The No-Limits Database47 High Speed Ingest Fast bulk load or stream data with real-time pipelines Memory Optimized Tables Ultra-low latency for transactions and analytics Disk Optimized Tables Petabyte scale analytics with compression and performance
- 48. Ecosystem High Speed Ingest Memory Optimized Rowstore Disk Optimized Columnstore Real-Time Data Messaging and Transforms Data Inputs BI Dashboards Kafka Spark Relational Hadoop Amazon S3 Bare Metal, Virtual Machines, Containers On-Premises, Cloud, As a Service Real-Time Applications Tableau Looker Microstrategy 48 Relational Key-Value Document Geospatial
- 49. Fast Data Ingestion ● Stream ingestion ● Fast parallel bulk loading ● Built-in Create Pipeline ● Transactional Consistency ● Exactly-Once Semantics ● Native integrations with Kafka, AWS S3, Azure Blob, HDFS 49
- 50. Instant Insights ● Scalable ANSI SQL ● Full ACID capabilities ● Support for JSON, Geospatial, and Full-Text Search ● Fast Query Vectorization and Compilation ● Extensibility with Stored Procedures, UDFs, UDAs 50
- 51. Thank You