Re: [vnrg] Logical vs. virtual

[Martin Stiemerling <Martin.Stiemerling at neclab.eu>]

Dear all,

I'm replying to this rather old email thread about "Logical vs. virtual", as I haven't found a compelling definition yet of what is logical and what is virtual. Not that it wasn't stated before, I have just a hard time to catch the definitions. 


After spending some time with reasoning (on my own and with a colleague) about logical and virtual, I see them orthogonal to each other, as the perception if something is logical or virtual may heavily dependent on the view. 

Logical seems to me the description of a thing (entity), in our case a (virtual) network. A virtual network logically consists out of virtual links and virtual routers (omitting switches, etc here). This would be the view from top (user of CN) to down. On the other hand, seen from the physical side, a virtual network logically consists out of instances on a physical link, instances on a router, etc.

Virtual on the other hand describes more the fact that you can use something which is actually not really existing, plus some more features as the ability to relocated, re-visitation, etc, probably going beyond the abilities of the lower (physical layer). An example would be direct link between two virtual routers, which in fact crosses multiple routers, i.e., you get the direct link which is actually not existing but constructed. 

To give an example from the computer world on virtual memory:
Virtual memory gives you something beyond the ability of real physical memory, e.g., continuous memory or memory beyond the real RAM by using space on a hard disk. Logically you have memory for application and the OS logically provides you with the means (real RAM or disk space) to have such a memory. 

Somebody (probably Joe) mentioned also that there is a mapping step from a logical level to the level below, e.g., mapping from need "x routers and links in a star topology" to the layer below (physical or virtual) in terms of "take VLAN X, router instance Z on router DS". 

The above text is for sure not ready to be casted in stone, but an attempt to get some textual definitions. Ideas of the text are not all mine, but include a condensed ideas of emails.

Open fire on my right now! :)

  Martin

martin.stiemerling at neclab.eu

NEC Laboratories Europe - Network Research Division
NEC Europe Limited | Registered Office: NEC House, 1 Victoria Road, London W3 6BL | Registered in England 2832014 


> -----Original Message-----
> From: vnrg-bounces at irtf.org [mailto:vnrg-bounces at irtf.org] On Behalf Of
> Joe Touch
> Sent: Thursday, July 15, 2010 6:51 PM
> To: Didier Colle
> Cc: vnrg at irtf.org
> Subject: Re: [vnrg] Logical vs. virtual
> 
> (again, speaking as an individual participant):
> 
> On 7/15/2010 9:36 AM, Didier Colle wrote:
> > Dear Joe, all,
> >
> >> I tend to think of a logical device as not really adding to the
> >> capabilities of a physical one, though, whereas a virtual one seems
> to
> >> (at least to me).
> >
> > Please explain. What capabilities are you referring to?
> > So far, I think most people on the list have expressed a virtual
> network
> > having the same or a subset of the capabilities of the network. Do
> you
> > agree with that? And remains that statement true for a virtual device
> > rather than virtual network?
> 
> A physical network can't be relocated while running. A virtual one can.
> A physical network doesn't ensure separation from other physical
> networks; a virtual one should, IMO.
> 
> >  From your previous post today, these capabilities are not apparent.
> >> 1.c. what is the characteristic behavior/capability of the
> >> resulting system?
> >>
> >> I think that the defining characteristic of a VN is that it allows
> an
> >> existing network to concurrently emulate another, distinct and
> >> separate network.
>  >
> > Is that a characteristic/capability of the virtual network (VN) or of
> > the underlying existing network?
> 
> The VN. Without the VN, no such emulation would occur, IMO.
> 
> >> Maybe these are really just two aspects of a single thing, i.e., I
> >> would differentiate between:
> >>
> >> - a device which does not map 1:1 to a physical entity
> >> (could be a part of one, a group of many, or a group of parts)
> >>
> >> - a device which provides supports virtual networking
> >>
> >> Since we call the latter "virtual networking" (the name of this RG),
> I
> >> would propose that the latter would be the virtual device, and the
> >> former be the logical one.
> >
> > Hmm... so the resulting question is: What is virtual networking? ;-)
> 
> I had proposed a definition - which is a network composed of tunnels.
> The rest, IMO, is largely in support of that one distinguishing
> characteristic.
> 
> >> I don't think of it as a 'specialized form', but rather the fact
> that
> >> a single virtual router CANNOT be mapped onto a single physical
> device
> >> with no other logical components.
> >>
> >> I.e., a virtual router needs a base router to connect to the links
> of
> >> the tunnel. A virtual host needs a base router (as well) for the
> same
> >> reason.
>  >
> > Really? In your previous post you said: "A virtual link is the
> easiest
> > to define - it is a tunnel over some existing network path, i.e.,
> with
> > an additional layer of encapsulation that is used solely for the VN,
> but
> > which is otherwise not needed. A virtual router forwards packets
> between
> > virtual links. " So, virtual link is encapsulated (e.g., by tagging
> with
> > a VLAN-ID?) to allow multiplexing: is this something the virtual
> router
> > cannot do (although invisible to the user/owner of the virtual
> router)?
> 
> A virtual router can encapsulate, but doesn't know what to do with the
> encapsulated packet, IMO. At some point the packet needs to go out a
> physical interface, and I don't think virtual routers own any physical
> interfaces.
> 
> > Do you need for that another "base" router?
> 
> No; that's what I needed the base router for in the first place.
> 
> Joe
> _______________________________________________
> vnrg mailing list
> vnrg at irtf.org
> https://www.irtf.org/mailman/listinfo/vnrg