In general, my interest is an estimation of possibility of integration of some features like usage of containers (light-weight virtualisation), checkpointing and persistency, applications migration / etc with modern microkernels. May be with verification (while this is a different story).
To do so I make a preliminary port of golang and related stuff onto of Genode to better understand related problems and as a base for containers support (while I am not a Go programmer, I never write any programs on it, just fix others ;). In this moment for me somehow obvious that to support things above I need a support in Genode some parts of generic virtualisation. Like namespaces based isolation (read: ability to have same names/id’s/etc in different domains for objects and anything provided by the Genode to user apps, together with additional related API). At least for app snapshotting, migration and persistency this is «the must». They are not so necessary for containers themselves, there are support of some platforms without it, as well without dedicated layered FS (unions and similar like auFS/btrfs/zfs/etc - while it is good to have it). Note: I suspect that having namespaces virtualisation on the kernel level will give some additional advantages for Genode even in terms of security. Same like happens in Linux with our proposals related to openvz/namespaces/userbeancounters (after some time it became clear that they are necessary for modern OS). This is relatively cheap in implementation and overhead points of view. Did you considers this option as a part of Genode future? Alexander > 23 дек. 2021 г., в 21:05, Norman Feske <[email protected]> > написал(а): > > Dear Genode community, > > it is the time of the year again to reflect and make plans for the > foreseeable future. Hereby, I'd like to kick off our traditional > brainstorming about Genode's road map for the year ahead of us. > _______________________________________________ Genode users mailing list [email protected] https://lists.genode.org/listinfo/users
