|
|
View Forum Thread
Anyone can view the forums, but you need to log in in order to post messages.
> Forum Home >
Announcements >
The new database blog
|
Tue, 11/Oct/2011 1:56 PM |
Robin
1138 Posts
|
Hi everyone.
I am writing here to update you about the status of development of the new database and communication system for Smedge. We are still early in actual implementation, but things are going very smoothly. We have the core database primitive system working, which allows storage, retrieval, and communication of any arbitrary data. This means that most version issues will be a thing of the past, as any version will be able to pack or extract any data without needing to know anything about how that data is used.
We have also implemented the first level of communication, successfully transmitting both "get" and "set" requests successfully from client to server and having them received and processed correctly.
With the current alpha code, with extra debugging code and no optimization, it was still able to process through 10,000 randomly generated messages in less than 15 seconds, only using about 30-40% of the cores on my 8 core Mac Pro (2008 vintage). This is a vast increase in performance over the older version, and we haven't done any optimization yet.
Currently, communication is moving forward with the second stage, returning information from the server back to the client, and then will be moving on to the synchronization system between the nodes. And database is moving forward with the fallback mechanism that will allow arbitrary overlays of defaults to be applied to data. This mechanism will allow behaviors to be customized at various levels automatically. For example, if you are getting a parameter from a job, that parameter may come from a platform override, or from an engine option or from the product settings, any of which you can customize for your pipeline, dynamically and system-wide, without restarting Smedge.
I will keep posting more here as we make more progress. Thank you for your support, and let us know if you have any questions!
-robin
|
|
|
|
Thu, 02/Feb/2012 3:48 AM |
Robin
1138 Posts
|
Hi everyone.
An update on the new database. It is well into implementation. I am working now to finish off Smedge 2012, which will be a major improvement on the current platform in performance and stability. But the new database is an essential next step in the process.
As far as functionality, most of the core system interaction is running. The routing and communications systems are implemented, and the system can automatically elect a master to coordinate things. Next stages are to start implementing the meta-data systems in order to be able to build the product data. Then begins the merging of the existing product control and monitoring code. At that point, things should be just about ready for preliminary testing. More features will then be implemented as it unfolds.
-robin
|
|
|
|
|
|
. |