Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Hi folks!
    This room is dedicated to discuss the End-point Masquerading Set of rules
    Sunil Mahendrakar
    @SunilMahendrakar
    Hey everyone! I'm a CS undergrad at National Institute of Technology, Rourkela, India. I'd love to work on this project (or IGDP) during GSoC this year. I was a GSoC'15 student developer for pagmo (an optimization toolbox maintained by ESA). Here are more details on the project. I am also an undergrad researcher (Network and Secured Computing Group) at my university. Are there any starter tasks for this project so that I can get acquainted with the code base?
    Prudhvi Dharmana
    @prudhvid
    Hello @vicente-gonzalez-ruiz I would love to work on EMS for the GSoc 2016(other options include virtual room). I went through the protocol pages. I'm currently thinking of going through the existing src code. Would you advice me to do anything else? I would be happy if you share any architectural desgin of the code that you've followed till now. Thanks
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Hi @SunilMahendrakar, welcome! We have tons of "tasks" to develop, but not "starter". Sorry. If you are interest in this idea (EMS), I can tell you that we want to achieve is that two or more peers in the same network (same NAT) can belongs to the same team. Now, in the current implementation this generates problems because the "private" peers see the public end-point of each other, when they should use the private ones.
    Hello @prudhvid, at this moment I can only help you with the objectives of this idea. If you want, start working on the implementation and I'll try to guide you. Thanks!
    Prudhvi Dharmana
    @prudhvid
    Thanks @vicente-gonzalez-ruiz I want to submit a proposal regarding this. Could you give me suggestions on things that I've to do before submitting the proposal on this?
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    @prudhvid, the most important thing in a proposal is to develop a feasible "roadmap". I means, to say what do you want to do, how and when.
    Prudhvi Dharmana
    @prudhvid
    Sure @vicente-gonzalez-ruiz . Will work on it Thanks :)
    Sunil Mahendrakar
    @SunilMahendrakar
    Hey! @vicente-gonzalez-ruiz thanks for the warm welcome. I'll send you the proposal soon.
    A.Ramya Keerthana
    @RamyaKeerthana
    hello...i am computer science undergrad..i want to work for this project for GSOC . I have basic skills on computer networks, sockets and C++.I have implented peer-peer file sharing protcol as a part of our academic project.Can any one guide me ... as in where to start ?
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Hi @RamyaKeerthana, welcome! EMS will be in charge of allowing more than one peer (belonging to the same team) behind the same NAT box. Please, read the P2PSP documentation available at GitHub and we will be glad to answer all the questions which surely will arise :-)
    Kevin Shi
    @kshi219
    Hello Vicente! I am a computer science student from Canada Interested in this project, I have just submitted a draft proposal via the GSOC website, I hope this is not too late, I was wondering if you could give it a quick look over, mainly I am not very confident that I've fully understood all the things that need to be done to fully address the problem. I have also sent the proposal to your email
    sajid hussain
    @sajjusajuu
    hey vincente.! I see that p2psp has already implemented NAT set of traversal rules(splitter_nts.py, peer_nts.py, monitor_nts.py) But to implement NTS the splitter has to know if a peer is behind NAT, which has to be implemented in EMS. How come NAT has been implemented without EMS? Thank u in advance.!
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    @kshi219, welcome! Please, revise your email.
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    @sajjusajuu, NTS enables the interchange of chunks between NATed peers, but only one peer/NAT is allowed. On the other hand, EMS enables the interchange of chunks between NATed peer, but peers must be behind the same NAT. As far as I remember, NTS handles all peers (NATed or not) the same way. Public and cone-NATed peers will use a fixed public end-point and the rest of peers, a varying public end-point, but in any case a test in run to check the peer is NATed or not. I think that this test should be implemented in EMS.
    mightyCelu
    @mightyCelu
    This message was deleted
    This message was deleted
    sajid hussain
    @sajjusajuu
    @vicente-gonzalez-ruiz Hello..! I have shared a draft for EMS. please comment!
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    @sajjusajuu done :-)
    sajid hussain
    @sajjusajuu
    Any improvements needed? or did I get wrong anywhere?
    @vicente-gonzalez-ruiz yes I want to add EMS features for NTS module too..!
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Not, I think that it is fine.
    (except for the question I did)
    My question is ...: will EMS inherit from NTS?
    sajid hussain
    @sajjusajuu
    yeah for the question.! I want to add ems functionality to NTS also. not by inheriting EMS from NTS but the other way.! NTS inheriting EMS.
    sajid hussain
    @sajjusajuu
    EMS inherits from DBS. Implementing this EMS is first part of my plan, and the present version of NTS module wants to solve the problem if peer are behind NATs( of various kinds) but it did not care if peers are behind same NATs . I think that we should modify the present NTS module so that it could have take care of that case too where peers are behind same NATs. For this I want to extend the NTS module from the EMS module created in 1st part.!
    Did I made myself clear or am I missing something?
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Hi @sajjusajuu, thanks for your comments. You should know that NTS is being translated to C++ and this moment. However, I think you could do your work. Thanks!
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Hi @kshi219. We would like to known your thoughts about building EMS on top of NTS (EMS-enabled peers should be behind a NAT box and therefore, should also run NTS).
    Kevin Shi
    @kshi219
    Hi @vicente-gonzalez-ruiz , I'll have to look into this a bit more (specifically the c++ version of NTS), but from researching my proposal the idea i got was that a lot of the specialized functionality (the various traversal techniques) offered by NTS is not relevant to EMS because EMS applies to the specific case of peers behind the same NAT router, (as opposed to different NAT routers in the case of NTS). Using the implementation outlined in the proposal it would seem that we can have EMS without it relying on NTS. At present I have not look at our NTS in too much detail, I should have a clearer idea once I do.
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Yes @kshi219, it its possible to inherit EMS directly from DBS. However, we are pretty sure that EMS will be used at the same time that NTS (in parallel). So, how this two classes will be connected?
    @jellysheep, what do you think?
    jellysheep
    @jellysheep
    Well, I think (as you, Vicente also suggest)
    ... that EMS will only be used when NTS is needed anyway (when peers are behind a NAT device), so it would make sense to derive the EMS classes from NTS. Kevin, if your EMS implementation does not need functionality provided by NTS, you could just use this inheritance anyway, and not use NTS-special functions in EMS.
    In other words, EMS inherits from NTS, but does not need (or rely on) its functionality.
    Would that be a possible option for you both?
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Yes, I think so. But, maybe, another possibility could be to put EMS between and DBS. However, I see more reasonable (less prone to final "surprises") to build EMS over NTS.
    Kevin Shi
    @kshi219
    Ah I see, I was thinking only about needs of inheriting functionality, not about when things will be used. I absolutely agree that it would be more reasonable to build EMS over NTS (even if EMS does not need NTS functionality) because the two will almost always be used concurrently.
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Fine! Well, I think that at this moment we should start using a stable version of the P2PSP core with the NTS functionality included. If I'm not wrong, in the master branch of the core NTS does not exist. @jellysheep, do you think that NTS is mature enough to be merged into the master?
    jellysheep
    @jellysheep
    Ok, great! And in any case that EMS should indeed need a NTS function, we can just use that.
    Yeah, the C++ NTS implementation should be quite stable now. Merging the NTS branches into core/master and p2psp-console/master should work fine.
    jellysheep
    @jellysheep
    Just added the respective PRs
    Vicente González Ruiz
    @vicente-gonzalez-ruiz
    Thanks @jellysheep.