April 13, 2011

What Starcounter may be up to

Starcounter seems to be offering an in-memory object-based/object-oriented/whatever short-request DBMS that also talks SQL. I haven’t been briefed at this point, and hence don’t have detail beyond what’s on their rather breathless web site. I’m guessing this isn’t an H-Store/VoltDB architecture, but rather something more like what Workday runs.

Most of the crunch I found on the Starcounter website (emphasis mine) is:

Let’s say that it is possible to make a database that is 10,000 times faster than what you use today. It would then be possible for your computer language objects to live inside the database from the very beginning. From the first { Customer a = new Customer(); }. The objects could live in the database, not as a copy, but as both database object and a Java or C# object at the same time. The database would transparently be your heap. The time it would take to save your object to the database would be reduced to nothing.

If such a database existed, you could say goodbye to caches and the duality of business objects, the database objects/rows and the complexity that follows. The speed would be amazing. Goodbye to time consuming scale-out solutions. Actually, you would be able to say good bye to the databases as you know them. You only need your simple objects.

Such a technology would be the ultimate NoSQL database. But what if the ultimate NoSQL database had SQL support, ACID, checkpoints and recovery and other enterprise features? Your pure, clean objects would then become the fastest and most powerful database in the world.

Beside that, other clues to what Starcounter is doing include references to Hibernate and to the declining cost of RAM.


7 Responses to “What Starcounter may be up to”

  1. Vlad Rodionov on April 13th, 2011 2:12 pm

    -1. Dead pool.

  2. Quotees should be briefed before quoters | Strategic Messaging on April 13th, 2011 2:47 pm

    […] I just blogged about a company pre-launch because the news wasn’t actually embargoed (their website was up) and the press was asking me for comment. Those details are unusual, but I’d guess that the majority of quotes I give to the press are about news I haven’t been briefed on. […]

  3. Patrick McFadin on April 13th, 2011 4:36 pm

    Looks very interesting. I can’t wait to see an actual product! Sounds a lot like Terracotta Ehcache for Hibernate with an integrated DBMS backend.

  4. Dan Koren on April 13th, 2011 10:36 pm

    Sounds like someone has rediscovered object
    databases. Anyone remember GemStone, Versant
    or Objectivity? Wow!

  5. Curt Monash on April 14th, 2011 12:01 am

    Actually, Intersystems Cache’ is quite the well-adopted object-oriented DBMS.

  6. Joachim Wester on May 5th, 2011 11:04 am

    @Dan Koren

    You are close. The idea is that if RAM is the database medium, the database may act as

    1) An OO database.
    2) An OO heap.
    3) A relational database where the class is the table and the instance is the row.

    The idea further is that securing changes is very fast when the disk arm is still or when using SSD.

  7. Dan Koren on May 19th, 2011 2:01 am

    to Joachim:
    I’m afraid I see nothing new compared to Versant 20 years ago — which seems roughly the time it takes most folks to forget what had been done already and start another rediscovery cycle.

Leave a Reply

Feed: DBMS (database management system), DW (data warehousing), BI (business intelligence), and analytics technology Subscribe to the Monash Research feed via RSS or email:


Search our blogs and white papers

Monash Research blogs

User consulting

Building a short list? Refining your strategic plan? We can help.

Vendor advisory

We tell vendors what's happening -- and, more important, what they should do about it.

Monash Research highlights

Learn about white papers, webcasts, and blog highlights, by RSS or email.