Thanks for your support Dawid, In the near future I'd like to investigate sctp (there's a nio sockets implementation in the jvm, under the com.sun.* namespace) it might be useful for multipath jeri endpoints, over wireless and 3G (IPv6) networks. It wouldn't be a candidate for public use, at least not until Oracle decided to make their sctp implementation public.
Cheers, Peter. ----- Original message ----- > While I don't have an immediate use-case for this (all our River code > runs in a very trusted environment), I just wanted to say that I think > this is very interesting and valuable, along with so many of your > efforts, Peter! > > kind regards, > Dawid > > On 26/02/2015 12:24, Peter wrote: > > Some things I've been working on recently: > > > > Proxy codebase permissions; I annotate jar files with a > > META-INF/permissions.list file containing the permissions required by > > the proxy. > > > > Clients can find out what permissions proxy's require and grant these > > dynamically at runtime. > > > > Reversal of responsibility for bootstrap proxy's, traditionally we ask > > a smart proxy for the bootstrap proxy: > > > > A bootstrap proxy interface that allows clients to lookup a bootstrap > > proxy and ask it for the smart proxy after authentication. This > > allows DownloadPermission to be granted dynamically after > > authenticating the bootstrap proxy. > > > > An entry for listing interfaces implemented by smart proxy's: helps us > > lookup the proxy we want. > > > > Looking up the bootstrap proxy first, provides the following benefits: > > > > 1. Allows logical comparisons to be made locally before downloading a > > smart proxy codebase. 2. Allows authentication to occure before > > codebase download. 3. DownloadPermission can be granted dynamically, > > rather than by configuration. 4. Codebase downloads and smart proxy > > unmarshalling can be performed lazily. > > > > I have these features in my local copy of River, is there wider > > interest? > > > > Regards, > > > > Peter. > > > >
