Difference between revisions of "Open projects and ideas"

From vsd
Jump to navigation Jump to search
Line 21: Line 21:
  
 
libslirp: projects convergence towards a unified library
 
libslirp: projects convergence towards a unified library
 +
 +
Cado/scado: extend to inherited capabilities?
 +
 +
hashdns has been designed to give authoritative replies but it does not forward
 +
requests. In a scenario where all IoT/IoTh have public addresses and names,
 +
hashdns works well, but we need either to implement a forwarding service or
 +
to study a configuration for an external forwarder (dnsmasq? bind?) to
 +
support hash based DNS resolutions for private networks.
 +
 +
hashdns and fqdndhcp use syncronous access to the DNS server, this could slow
 +
down the service in case of bursts of requests. Implementations based on queues
 +
and asynchronous mapping of requests/replies should perform better.
 +
 +
fqdndhcp lease time is a constant, should be configurable
 +
 +
vxvde/vxvdex: BPF to filter spurious multicast self packets.
 +
 +
Performance analysis of vxvde using ....
 +
 +
exotcp
 +
 +
picotcp pycotcp
  
 
== Partial Virtual MAchines ==
 
== Partial Virtual MAchines ==
  
 
Purelibc library: update or reimplement?
 
Purelibc library: update or reimplement?

Revision as of 16:08, 29 March 2017

VirtualSquare general topics

Packetize new versions of our tools and libraries. (Giulia)

Creation of a new tutorial image.

Renew the communication means (update our wiki? New attention-catching tools are available?)

Virtual Networking

open question: vdens/vdestack is based on user namespaces... It is the case to port all the tools on user namespaces? Are them so widely adopted that the general publisc will use the new tools.

vdetelweb: new design (generalization of dynamic management of web pages and command auto-completion based on help messages)

vdetelweb: re-design vdetelweb on vde-stack.

full integration in the v2 architecture of picotcp/pycotcp.

VDE-4 switch, wirefilter...

libslirp: projects convergence towards a unified library

Cado/scado: extend to inherited capabilities?

hashdns has been designed to give authoritative replies but it does not forward requests. In a scenario where all IoT/IoTh have public addresses and names, hashdns works well, but we need either to implement a forwarding service or to study a configuration for an external forwarder (dnsmasq? bind?) to support hash based DNS resolutions for private networks.

hashdns and fqdndhcp use syncronous access to the DNS server, this could slow down the service in case of bursts of requests. Implementations based on queues and asynchronous mapping of requests/replies should perform better.

fqdndhcp lease time is a constant, should be configurable

vxvde/vxvdex: BPF to filter spurious multicast self packets.

Performance analysis of vxvde using ....

exotcp

picotcp pycotcp

Partial Virtual MAchines

Purelibc library: update or reimplement?