August 2nd Status Update
Since my last update I've created some significant developments around MckoiDDB (the new name of the software);
- The Mckoi distributed network stack is complete and fairly well tested.
- I have implemented a Consensus specification for managing how concurrent transaction proposals are merged during a commit.
- A simple (but very useful) database implementation which supports files, tables, indexes, and offers familiar transaction semantics has been added.
- Blocks now get compressed using gzip.
- I've added some basic administration tools.
- The content management software running this website has been ported to a distributed installation of MckoiDDB under Jetty.
I now just need to package out the binaries and source code, and complete the documentation, and we'll be ready to go.
Comments
Very good news, we will be waiting for the 1st release!
Regards,
Martin Cordova
Regards,
Martin Cordova
Toby, my friend... it's great news! but it would be great to have a snapshot of what you're working on: probably setting up a public Git or SVN would make us understand and possibly contribute! :)
Unfortunately I still have some licensing issues to clear up before I can release any code.
may I ask you if you'rve taken in consideration the ACID transaction model for the commit?
PS. I understand the problems related to the release of code partially under license restrictions: it's though!
PS. I understand the problems related to the release of code partially under license restrictions: it's though!
The ACID properties are all there. MckoiDDB follows a very strict isolation model and is strongly consistent, much like Mckoi SQL was.
Wow! Terrific news! I look forward to put my hands on the binaries of McKoiDDB!
I understood that SQL language (and McKoi dialect) is not used anymore and the object-oriented approach it's preferred now: is that true? If not, which version of SQL is it implemented (SQL-92 or SQL-99, which includes the object notations)?
Have you included a system for the programmability of procedures and functions (such as PL/SQL or T-SQL)?
Have you included a system for the programmability of procedures and functions (such as PL/SQL or T-SQL)?
There is no specific database model implemented in the core MckoiDDB software, except an API called the Simple Database API. The Simple Database API implements a basic set of useful persistent structures that you use through Java. These objects often implement some form of the Java collections API.
We aren't getting involved with SQL standards just yet, but there will be movement towards that. Right now I am more interested in providing a solid framework on which those data models can be built in a strictly distributed environment. It will be possible to implement graph structured databases, relational databases, file systems, or any data model you want using MckoiDDB.
We aren't getting involved with SQL standards just yet, but there will be movement towards that. Right now I am more interested in providing a solid framework on which those data models can be built in a strictly distributed environment. It will be possible to implement graph structured databases, relational databases, file systems, or any data model you want using MckoiDDB.
I've been following your work for years and always love reading about what you are up to next... I imagine MckoiDDB will be a solid framework with tremendous potential when you are done. Over the years, I've used Berkeley DB in our project and have had good success with it. We've implemented this as a layer of caching to improve performance rather than hitting a RDBMS database for everything. I'm anxious to see how you've implemented certain things. I'm assuming since the database is mostly programmatic database, you will release this as a framework for building other kinds of distributed databases such as GIS databases, XML databases, etc. I'm sure the possibilities are endless.