On Sun, 02 Sep 2007 18:57:30 -0400 Ken Mandelberg <[EMAIL PROTECTED]> wrote:
> Bob Doolittle <[EMAIL PROTECTED]> Writes: > > >We're doing a lot of things, but we can't do > >everything. There are features of Sun's products > >that are of little interest to the vast majority > >of SRSS users. Would it make sense to focus on > >them at the expense of features which are known to > >be more interesting to the community, even if, for > >example, they are unrelated to the rest of Sun's > >products (e.g. Windows Connector). > > > >We have to prioritize and offer the features we > >judge best for our customers with the resources we > >have available. Good management and > >prioritization is what keeps costs from > >skyrocketing and making products unaffordable. > > > Nearly everything at Sun but Sunray has been open sourced, which allows > the community to expand the efforts of the internal developers in > multiple directions. With so many things needed and so few internal > resources, I really can't understand why Sunray is the hold out. Yes. Don't we all remember the hype around <http://www.sun.com/smi/Press/sunflash/2005-11/sunflash.20051130.1.xml> ? I thought about this for a while and discussed it with some friends. I am a scientist and sysadmin without any business background, so please be patient if I got everything wrong or missed the obvious, but here are some of my thoughts: I am afraid the present development model of SRSS for Linux will not work in the long run. People simply don't like to be stuck with one or two supported distributions (even worse: specific releases of these distributions). And SRSS is a beast. It interferes with the network configuration, the PAM authentication system, provides kernel modules meddling with USB and audio, installs its own LDAP database, brings its own X server, starts several daemons... With an open source SRSS, the community would have at least some chance to adopt it to different and recent distributions. One could at least dare to ask on the Linux kernel mailing list "Why is this [EMAIL PROTECTED] kernel module no longer working with recent kernels?" But I am afraid there is some reason we don't hear anything about an open source future of SRSS. A full open source SRSS, including an open specification of the ALP protocol, could have several consequences,e.g.,: - People could write a "software Sun Ray client". I think this would be cool and would even help to spread out the SunRay technology. But Sun already has the "Secure global desktop" aimed at similar fields of application. - Vendors could produce SunRay-compatible clients without a licence. I don't know if this is a real danger to Sun. I doubt that one can make them really better or cheaper than Sun's devices. I also don't know if an open source software licence could oprevent this by some legal clauses. - The ALP protocol includes a great deal of ideas and tricks for an optimal compression of the network traffic in order work with quite different network bandwiths giving always an optimal "response feeling" to the user. This could be of interest to a lot of other vendors, even to Microsoft's RDP or Citrix' ICA protocol. Perhaps this is the main point. - Another topic are 3th party secrets which already are or may come into SRSS: If the Sun Ray technology wants to stay and prosper, it must improve on the multimedia side. Video streaming, 3D acceleration, better audio quality and so on. I am sure developers at Sun are working hard in this direction. And here we are really entering the dark realm of patents, vendor secrets, NDAs, proprietary interfaces and lawyers. I am afraid this is the true reason why we don't hear anything about an open source future for SRSS. SRSS 10.0 will have greatly improved video performance but it will be "protected" by dozens of patents and NDAs. So what can we hope for? Will Sun make at least large parts of SRSS open source, in order to help porting it to other Linux distros? Anyhow, besides any Solaris vs Linux or open vs closed discussion: It would be a good thing to see Sun commited to some concrete roadmap for the future of SRSS. And a final note: I could live with an Solaris server for SRSS and a Linux server providing desktop and applications (they could even run, virtualized, on the same piece of hardware) if I knew how to redirect USB and audio devices to the other server. A kind of "Sun ray connector for Unix/Linux" providing this feature would be great. One could extend it to a Xen-based or a zones/brandz based "VDA for Solaris/Linux" providing a virtual machine for every user. Regards, Meik -- Meik Hellmund Mathematisches Institut, Uni Leipzig e-mail: [EMAIL PROTECTED] http://www.math.uni-leipzig.de/~hellmund _______________________________________________ SunRay-Users mailing list [email protected] http://www.filibeto.org/mailman/listinfo/sunray-users
