0%

Book Description

The database pendulum is in full swing. Ten years ago, web-scale companies began moving away from proprietary relational databases to handle big data use cases with NoSQL and Hadoop. Now, for a variety of reasons, the pendulum is swinging back toward SQL-based solutions. What many companies really want is a system that can handle all of their operational, OLTP, BI, and analytic workloads. Could such an all-in-one database exist?

This O’Reilly report examines this quest for database nirvana, or what Gartner recently dubbed Hybrid Transaction/Analytical Processing (HTAP). Author Rohit Jain takes an in-depth look at the possibilities and the challenges for companies that long for a single query engine to rule them all.

With this report, you’ll explore:

  • The challenges of having one query engine support operational, BI, and analytical workloads
  • Efforts to produce a query engine that supports multiple storage engines
  • Attempts to support multiple data models with the same query engine
  • Why an HTAP database engine needs to provide enterprise-caliber capabilities, including high availability, security, and manageability
  • How to assess various options for meeting workload requirements with one database engine, or a combination of query and storage engines

Table of Contents

  1. In Search of Database Nirvana
    1. The Swinging Database Pendulum
    2. HTAP Workloads: Operational versus Analytical
    3. Query versus Storage Engine
    4. Challenge: A Single Query Engine for All Workloads
      1. Data Structure—Key Support, Clustering, Partitioning
      2. Statistics
      3. Predicates on Nonleading Key Columns or Nonkey Columns
      4. Indexes and Materialized Views
      5. Degree of Parallelism
      6. Reducing the Search Space
      7. Join Type
      8. Data Flow and Access
      9. Mixed Workload
      10. Streaming
      11. Feature Support
    5. Challenge: Supporting Multiple Storage Engines
      1. Statistics
      2. Key Structure
      3. Partitioning
      4. Data Type Support
      5. Projection and Selection
      6. Extensibility
      7. Security Enforcement
      8. Transaction Management
      9. Metadata Support
      10. Performance, Scale, and Concurrency Considerations
      11. Error Handling
      12. Other Operational Aspects
    6. Challenge: Same Data Model for All Workloads
    7. Challenge: Enterprise-Caliber Capabilities
      1. High Availability
      2. Security
      3. Manageability
    8. Assessing HTAP Options
      1. Capabilities of the Query Engine
      2. Integration Between the Query and Storage Engines
      3. Data Model Support
      4. Enterprise-Caliber Capabilities
    9. Conclusion
3.128.202.221