Re: [tipc-discussion] Non-fully meshed network and secondary nodes

[Stephens, A. <all...@wi...>]

Hi Ian:

The secondary/slave node concept was proposed in the initial TIPC
specification, but never fully developed.  There is prototype code in
the TIPC 1.6 code in the Linux kernel, but it is incomplete and untested
and no further development in this area is anticipated.

Fortunately, it is still possible to construct a TIPC network that is
not fully meshed by taking advantage of the hierarchical network support
(aka multi-cluster/multi-zone support) that was added in TIPC 1.7.  For
example, you could implement your A-B-C-D example by treating A-B and
C-D as separate clusters that have an inter-cluster link between B and
C; that is, A-B=C-D.

More detailed information on hierarchical network support is provided in
the TIPC 1.7 Programmer's Guide and TIPC 1.7 User's Guide, which can be
found at http://tipc.sourceforge.net/documentation.html.  Be sure to
take note of any limitations that may make it unsuitable for your
particular system.  The most significant of these limitations are:

1) The distance between any 2 nodes in the network must be no more than
7 hops.  For example, A-B=C-D#E-F=G-H has nodes A and H in two different
zones at the maximum possible distance.

2) TIPC's multicasting capability is only supported between nodes in the
same cluster.  Thus, it isn't possible to directly multicast a message
to nodes in other clusters or zones.  (You could, however, unicast a
message to an agent in another cluster or zone, and have it do the
multicasting ... but you'd have to implement the agent yourself.)

Regards,
Al

> -----Original Message-----
> From: Ian Huang [mailto:ih...@qu...] 
> Sent: Thursday, February 19, 2009 7:35 PM
> To: tip...@li...
> Subject: [tipc-discussion] Non-fully meshed network and 
> secondary nodes
> 
> Hi,
> 
> I'm looking at using TIPC with a number of nodes linked 
> together with point-to-point links, so it is not possible for 
> the network to be fully-meshed. I came across the concept of 
> secondary/slave nodes, but I cannot find any reference to it 
> in the latest code base although the code in the Linux kernel 
> seems to have quite a few references to it. Are 
> secondary/slave nodes supported?
> 
> If not, how can I build a TIPC network that looks like the 
> following and have all the nodes talking to each other?
> 
> A-B-C-D
> 
> Thanks,
> -Ian
> 
> 
> --------------------------------------------------------------
> ----------------
> Open Source Business Conference (OSBC), March 24-25, 2009, 
> San Francisco, CA -OSBC tackles the biggest issue in open 
> source: Open Sourcing the Enterprise -Strategies to boost 
> innovation and cut costs with open source participation 
> -Receive a $600 discount off the registration fee with the 
> source code: SFAD http://p.sf.net/sfu/XcvMzF8H 
> _______________________________________________
> tipc-discussion mailing list
> tip...@li...
> https://lists.sourceforge.net/lists/listinfo/tipc-discussion
>