Hi, Great job putting together this roadmap!
Regarding 1) you may find this work relevant: http://arxiv.org/abs/1503.00626v1 Regarding 2) the semi-streaming model allows polylog(n) passes. If you allow only one pass, it's a streaming model. In addition, I think models such as the turnstile/fully dynamic (with addition and deletion) would be more appropriate than the incremental one. Regarding 5) (when combined with 2) there are some good recent algorithms for partitioning that you might want to look at. http://dl.acm.org/citation.cfm?id=2556213 http://dl.acm.org/citation.cfm?id=2339722 Cheers, -- Gianmarco On 18 May 2015 at 12:42, Vasiliki Kalavri <vasilikikala...@gmail.com> wrote: > Hello everyone, > > thank you Andra for putting together this nice document and for sharing! > > Gelly is already in a state that matches the basic functionality of > existing distributed graph libraries. We're very happy for this and also > very grateful to everyone who has contributed so far! > > We are also very excited to keep on extending Gelly and our vision is to > turn it into something much more than "yet another library for running > PageRank-like algorithms". The document that Andra has shared contains > ongoing work and our ideas on mid- and long-term plans, in order to make > this happen :) > > We would love to get the community's feedback on this, so please feel free > to comment on the doc or discuss individual points here. > > Cheers, > Vasia. > > On 18 May 2015 at 11:01, Andra Lungu <lungu.an...@gmail.com> wrote: > > > Dear Squirrels, > > > > The following document is intended to give you an insight on where > Flink's > > Graph API, Gelly, is at the moment and especially on the direction in > which > > it's heading in the near future. > > > > > > > https://docs.google.com/document/d/11ovK1xZZTl3Qinw6Ud1rqpaDhTvRHDv8PWBzcHITFkM/edit?usp=sharing > > > > We would love to hear about your comments/ suggestions/ additions to the > > wish-list :) > > > > Cheers, > > Andra > > >
