1 / 7

Grand Challenges in Testing Data-intensive Computing Systems DBTest Workshop 2010

Leo Giakoumakis, Microsoft SQL Server. Grand Challenges in Testing Data-intensive Computing Systems DBTest Workshop 2010. Terminology. Testing is: Ensuring that the system is built as designed Ensuring customer requirements are met Finding bugs

gallia
Télécharger la présentation

Grand Challenges in Testing Data-intensive Computing Systems DBTest Workshop 2010

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Leo Giakoumakis, Microsoft SQL Server Grand Challenges in Testing Data-intensive Computing SystemsDBTest Workshop 2010

  2. Terminology • Testing is: • Ensuring that the system is built as designed • Ensuring customer requirements are met • Finding bugs • Quantifying, tuning and deciding engineering tradeoffs • Providing insurance • Providing information: learning • Data management systems: • Broad term: includes any of SQL Server, Stream Insight, SAP, Big Table, Hadoop, etc.

  3. Data management & Testing • Data management field • More than 30 years of research and development • Built on strong foundations • Enormous body of published work by academia and industry • Today anyone can build a RDBMS, the “text-book” has been written! • MySQL, Postgress, and other code bases in the academia are available • The field of testing data management systems • Test engineering is an immature discipline • Very little information is shared • The state of the art is yet to be defined • It’s time to start writing the text-book on testing data management systems!

  4. What makes testing challenging • Size and Cost • Testing at realistic scale and size is hard • Testing the size and richness of the programming surface • The cost of testing an increment of the feature-set is often a function of the entire feature-set • Test code bases are becoming large, hard to manage • Test immortal test case problem! • Understanding coverage • Knowing when you have done enough testing is hard • Code coverage is not enough; you need state, workload, scenario coverage • Complexity • Testing is multidisciplinary: language compiler, optimization, operating system, etc. • Complexity increases with appliances, distributed systems, the Cloud

  5. Areas in need of solutions 1/2 • Workloads/benchmarks • Standard benchmarks are: performance oriented, simple, linear • We need large, mixed workloads with un-steady states and built-in failures • Methodology and metrics for workload characterization • Important: this is what most researchers rely on to evaluate their ideas • Test architecture and test reuse • What are the abstract primitives that would apply to testing most systems? • A standard blueprint for testing data management systems? • A standard suite of test methodologies? • The industry is moving towards “multiple engines” • Large data • Data expansion and data reduction

  6. Areas in need of solutions 2/2 • Query optimization quality • Defining metrics for QO quality is hard • Also, measuring quality over time and over code changes • How can you compare optimizersof two competitive systems? • Proving query results correctness for complex queries over large data • The Cloud • Distributed data processing/storage at a large scale • Failure: from unlikely to certain • Failure conditions: harder to reproduce • Testing needs to takes place “in production” too • SLA testing

  7. Thank you!

More Related