How many TB/PB is the largest heavydb database in use in the production environment so far?

Follow

jieguo

• November 03, 2023 03:11

Three questions,thanks:
1.How many TB/PB is the largest heavydb database in use in the production environment so far?
2.What is the maximum number of concurrent users supported?
3.Is the heavydb database suitable for OLTP, OLAP or HTAP?

0

• 

  Candido Dessanti

  • November 12, 2023 15:23

  Hi @jieguo,

  We don't have access to such specific data, and in any case, we wouldn't disclose information about our paying customers. What I can share is that our database is often utilized alongside other analytics solutions such as Snowflake, Big Query, Vertica, or Teradata due to its superior performance.

  While the topic can be complex, in simplified terms, our system processes GPU queries serially and CPU queries concurrently. The number of queries that can run concurrently is governed by the num-executors parameter, which defaults to 4. By concurrently, I mean wqueries actively wunning within the same nanosecond. These limits can be overcome by employing various distributed configurations, such as an example launching more instances of the engine on the same database.

  Our database is specifically designed for running analytical queries and exclusively utilizes a columnar store. As a result, it is not suitable for OLTP (Online Transaction Processing), a use case that has never been within our scope. Unlike the HTAP (Hybrid Transactional/Analytical Processing) approach, which requires an engine to support different data stores, our focus is on a unified architecture.

  In contrast to databases like Oracle, which employs a row-based store for OLTP and a columnar store for in-memory processing (not to mention hybrid columnar storage, which is often deemed less effective), our system does not follow this hybrid model. While the term OLAP ) might seem somewhat limiting to describe our offering, it remains the most accurate standard definition for the capabilities our database is optimized for.

  
  Best regards,Candido .

   

  1

  Comment actions Permalink
• 

  jieguo

  • November 13, 2023 01:46

  Thanks!

  We don't need to know customer information, we just want to know how big the heavydb database is currently running.

   

  So we have an evaluation of database reference selection.

  0

  Comment actions Permalink
• 

  Candido Dessanti

  • November 14, 2023 15:10

  Hi,

  Database sizes exhibit significant variability, influenced by factors such as usage patterns and the desired performance during dashboard navigation, particularly when used in conjunction with Heavy Immerse.

  In our ecosystem, data within the database is typically fully utilized for processing purposes. As a standard practice, old data is either purged or moved out of the database rather than being retained for future use. It's noteworthy that some of our customers load substantial amounts of data, exceeding 3 billion records, into multiple tables on a daily basis.

  This dynamic nature of data handling ensures optimal performance and resource utilization for our customers.

  If you have any further questions or need additional information, feel free to reach out.

  Best regards,
  Candido

  0

  Comment actions Permalink

Please sign in to leave a comment.