2013, we are still here. which means we march on. We started off last year, looking to ship libssh along with the code, as well as replace the older scripts we were using to start the code, by functions inside the code. The libssh server hit a few bottlenecks which has put that on the back burner for now while we beef up other important areas. The code does however use pipes and does not require the scripts to function.
Along with the changes, is a heavy rework and better multi-threading performance, Documentation is important. We now use Doxygen. A direct link to the code documentation, which can also be linked through from the Documentation menu. of course, we also created and released our much loved, Introduction video which seeks to introduce the concept and its usage. Very happy with the postive feedback, thank you :) A major part of the latter half of the year was spent in two critical areas which need improving.
- Sandboxing, LXC is our choice, and work is still underway to complete the integration.
- Switching to Apache Thrift, The internal messaging system for ebrainpool was long overdue for replacement. This will aid in both the immediate & in the future roadmap. Work is underway here too.
With these integrated, it will mark another huge step in the evolution of the code base. We can begin to unleash the feature sets and functionality which we have in mind. Some of it has been discussed, some of it still needs to be sifted through, and some still needs to be thought through, but this will be a good starting point. 2013 is going to be great year. Towards the end of the year, we had the chance to present this technology at OSDC 2012 (open source developers conference) in Sydney, Australia. Hopefully we got more people interested, time will tell :) We also have a 3d animation, visualising ebrainpool in action which was created for us. the First one is can be viewed in the link, we will be updating with an improved variant, which contains the entire sequence, along with explainations & hints. Roadmap for 2013 ?
We will be posting a more detailed post, but in brief, we want to streamline the development process more and get a release out, in binary form, as a deb & rpm which we hope will lead to more people using & evaluating this, find more bugs and giving us more to fix :) At the current time, the release is pending the completion of the sandboxing & thrift codebases and their merge in the main codebase. this is what we are thinking (v 0.5 & 0.7 are the important ones, the rest will be reactionary and could change)
- v0.5 : with thrift messaging system, cleanup of the current protocol in use, and a solid foundation for future features.
Looking towards Mid Februrary
- v0.7 : with sandboxing :) YEAH, this is the fun one :) Every application started, will have a common sandboxed area, More security, always a good thing :)
Looking towards end Februrary, early March -- followed by our initial binary release, in deb & rpm forms, ready for download and install, and wont require code compilation as it is currently.
- v0.8 : Some more settings, to overcomes issues which are proving to be annoying in real-world use. more details later.
- v0.9 : fuse integration, remote & local mounting of filesystems
- v0.9 -> v1.0, multiple interations : test with multi-computers, hopefully larger mesh ? Stress test, memory test etc..
Future directions, to be considered:
- load balacing, detect & distibute calls amongst best available clients, useful for closed internal networks
- evaluate architecture and introduce a plugin system, to allow integration into desktop etc.
- maybe separate out code as daemon & interface layer?
- evaluate a system to perhaps allow digital currency usage (generation / consumption) for those who wish to use it in that way. This needs to be non-compulsory. Another can of worms here :) Lets go fishing.
Thats all for now, more fun to follow. oh -- We are considering attending Battlemesh this year wink *wink ;)