The MPLS-OPS Archive

["Adrian Farrel" <adrian@olddog.co.uk>]

Sylvia,

 

One "detail" you hide is switching resources.

Labels may be sparse resources and switching a fat pipe only takes one label (or one slot in your matrix).

 

This get more important (but this is not the only reason for the function) when you change granularity of switching commodity. For example, if I want to send some small packet LSPs down a WDM fiber I could:

- map each LSP to a lambda, but I would quickly run out of lambdas

- wrap all of the packet LSPs in a lambda and demux the packets at

   each router, but this is inefficient at the router (which might not even

   be able to handle that number of packets). Note that in this case

   you are sort of building a single hop hierarchy.

 

Now, in your figure, the topology is linear and here your main saving is in the detailed switching and processing at transit routers. But even then, several parallel, tiny LSPs (maybe only 1 kbit) are hard to dynamically fit into an OC-48 switching network without some form of hierarchy or huge bandwidth wastage.

 

This may become clearer looking at a topology such as:

 

I1          E1

  \        /

I2-T1-T2-T3-E2

  /        \

I3          E3

 

and you wish to operate three LSPs I1-E1, I2-E2, I3-E3 then you can save switching resources (labels) on transit routers by building a tunnel from T1 to T3 and placing the three LSPs inside.

 

Another detail that can be saved by the use of hierarchical LSPs is complexity in the traffic engineering database. Consider a packet network where the routers are connected over a TDM network. The packet network wishes to believe that its routers are adjacent even though they are connected by many hops through the TDM network. To do this, you operate the TDM network on its own and establish LSPs between its edges - these LSPs are presented as TE links (FAs) to the packet network (think of virtual connections).

 

Cheers,

Adrian

 

----- Original Message -----

From: "Spice Sylvia" <falsesylvia@yahoo.co.uk>

To: <mpls-ops@mplsrc.com>

Sent: Tuesday, April 12, 2005 10:40 AM

Subject: Re: [MPLS-OPS]: Hierarchical LSPs

> Hi Amit,
> Please see inline
> -->
> --- Amit 
70405 <AmitG@huawei.com> wrote:
> 
> Hi Sylvia,
> >    FA is basically a higher 
bandwidth LSP.
> >    On top of this LSP you can  
combine(nesting) one
> > or more lesser banwidth LSP's.
> 
>    FA provides a kind of abstraction, that is a link
> 
> between FA ingress and egress.
> >    Thus hiding 
the details of that particular
> > portion of the network.
> > 
> --> and what "details" does one hide?
> If it is "link 
identifier" they are link local, if
> they are "site identifiers" they are 
site local, if
> they are end point identifiers, they have to be
> 
unique, if the problem is the "table size" you could
> do:
> 
N1---N2---N3--N7--N4--N5--N6
>      
|                 
|
>      N8-----N10-------N11
> 
> and 
advertise reachability to N6 via the N8 N10 path
> to N2 so that 
N3-->N5 is not used and hence reduce the
> table size. 
> So what 
is the exact details one is "hiding" and why
> is one hiding the 
same?
> 
> 
> >    There are many scenarios 
where we need this kind
> > of nesting.
> > 
> 
>    One of the scenario is Inter-area or Inter-As LSP
> 
> setup.
> >    Here an FA is setup in a single area 
and is used
> > by a LSP from another area that need to 
traverse
> > this FA LSP area.
> >    Thus the 
LSP will be setup without knowing the
> > details of the FA LSP 
area.
> > 
> ---> but that is equally valid for above, and if 
i do
> need to setup a constraint based LSP, each
> intermediate 
node which has reachability to the end
> point will have to consider all 
alternate paths, In
> other words, why dont I simply "knock" off those 
boxes
> in the middle and do full mesh :) /directly connect N1
> to 
N5?
> 
> 
> >  
> >    
Secondly in GMPLS.
> >    Say the backbone is a TDM or a 
Optical network.
> >    And access network is a PSC 
network.
> >    Bandwidth allocation in TDM or Optical 
will be in
> > higher chunk, than what it is in packet.
> 
> as far as my knowledge goes, TDM slots can be
> allocated per 
call, that is the bases of how existing
> PSTN/Digital Networks work 
today.
> 
> >    So in this case a FA is setup in 
the Backbone,
> > and multiple LSP's from access will start setting 
up
> > over this FA.
> > 
> 
> That FA is the 
"wire capacity" in my case, can you
> give me one reason where the 
"stacking/hierarchy
> helps". 
> 
> I would prefer practical 
examples, and would prefer if
> it was assumed every device in the middle 
has a
> control plane.
> 
> >    I hope it 
helps.
> > Regards,
> > Amit.
> 
> hope that 
clarifies my question
> -brgds
> Sylvia 
> > 
> 
>    
> >    
> > 
> 
> ----- Original Message -----
> > From: Spice Sylvia <falsesylvia@yahoo.co.uk>
> 
> Date: Tuesday, April 12, 2005 4:43 pm
> > Subject: [MPLS-OPS]: 
Hierarchical LSPs
> > 
> > > Hi,
> > > 
> 
> > Can someone please help me understand why the
> > > 
"capacity" of the wire itself cannot be treated as
> > the
> > 
> outer most constrain, rather than making a
> > > hierarchical 
association.
> > > 
> > > In other words:
> > 
> 
> > > N1---N2----N3---N7--N4---N5--N6
> > > 
> > > I see that folks define a FA from N3 to N4, inside
> 
> > which another FEC from N2 to N5 is created and
> > 
another
> > > N1 to N6 hierarchy is put inside it.
> > > 
> > > Howevr in all cases the CAC/setup is done from N1
> 
> to
> > > N6.
> > > 
> > > Is there a 
reason why the FEC hierarchy is built?
> > Why
> > > does 
the "capacity of the wire" not act as the FEC
> > > itself?
> 
> > 
> > > Is "maintaining the state" of each inner LSP 
on
> > > intermediate node a problem?
> > > 
> 
> > -brgds
> > > Sylvia
> > > 
> > > 
> > > Send instant messages to your online friends 
> > 
> http://uk.messenger.yahoo.com 
> > > 
> > > -------
> > > The MPLS-OPS 
Mailing List
> > > Subscribe/Unsubscribe: 
> > http://www.mplsrc.com/mplsops.shtml
> > > Archive:
> > http://www.mplsrc.com/mpls-ops_archive.shtml
> > > 
> > > MPLScon 2005 - May 16-19, 
NYC, NY
> > > http://www.mplscon.com/
> > > 
> > 
> > -------
> > The 
MPLS-OPS Mailing List
> > Subscribe/Unsubscribe: 
> > 
http://www.mplsrc.com/mplsops.shtml
> > Archive:
> > http://www.mplsrc.com/mpls-ops_archive.shtml
> > 
> > MPLScon 2005 - May 16-19, NYC, 
NY
> > http://www.mplscon.com/
> > 
> 
> Send instant messages to your online friends http://uk.messenger.yahoo.com 
> 
> -------
> The MPLS-OPS Mailing List
> 
Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
> Archive: http://www.mplsrc.com/mpls-ops_archive.shtml
> 
> MPLScon 2005 - May 16-19, NYC, NY
> 
http://www.mplscon.com/
> 
>