[vnrg] Logical vs. virtual

[Didier Colle <didier.colle at intec.UGent.be>]

Dear all,

Joe raised the issue of logical vs. virtual.
I believe it is worth spending a separate thread on clarifying this point.

At the bottom of this email is my proposal.

Joe stated:
> Note that nothing about these definitions specifies a boundary, i.e., 
> inside a single machine, etc. I don't think those boundaries are 
> meaningful in the base Internet anyway.
> I.e., IMO, virtual has nothing per se to do with "logical". I.e., a 
> set of devices on a network that source/sink packets with a single 
> network address act as a single logical host. That's not 'virtual' to 
> me, though most 'virtual' things tend to be logical, not all logical 
> things are virtual. 
1) Should I interprete this as logical referring to physical boundaries?
2) This does not define what virtual means or how it differs from 
logical --> we probably need to formally define virtual inside the vnrg 
to know what we are speaking about exactly.
3) I agree with "not all logical things are virtual". However, "MOST 
virtual things TEND to be logical" --> are there any counter examples in 
which virtual things are not logical?

This example also reminded me of the Virtual Router Redundancy Protocol 
(VRRP): see RFC3768.
Not sure what made the authors to call it virtual rather than logical.

According to Juniper (sorry for mentioning company names...): 
http://www.juniper.net/techpubs/software/junos/junos74/swconfig74-routing/html/logical-router-overview2.html
>     Logical Routers and Virtual Routers
>
> A virtual router is not the same as a logical router. A virtual router 
> is a type of simplified routing instance that has a single routing 
> table. A logical router is a partition of a physical router and can 
> contain multiple routing instances and routing tables. For example, a 
> logical router can contain multiple virtual router routing instances.

Virtual memory handling in computer systems: 
http://en.wikipedia.org/wiki/Virtual_memory#Segmented_virtual_memory
>       Page tables
>
> Almost all implementations use page tables 
> <http://en.wikipedia.org/wiki/Page_table> to translate the virtual 
> addresses seen by the application program into physical addresses 
> <http://en.wikipedia.org/wiki/Physical_address> (also referred to as 
> "real addresses") used by the hardware to process instructions. Each 
> entry in the page table contains a mapping for a virtual page to 
> either the real memory address at which the page is stored, or an 
> indicator that the page is currently held in a disk file. (Although 
> most do, some systems may not support use of a disk file for virtual 
> memory.)
>
> Systems can have one page table for the whole system or a separate 
> page table for each application. If there is only one, different 
> applications which are running at the same time 
> <http://en.wikipedia.org/wiki/Multiprogramming> share a single virtual 
> address space, i.e. they use different parts of a single range of 
> virtual addresses. Systems which use multiple page tables provide 
> multiple virtual address spaces—concurrent applications think they are 
> using the same range of virtual addresses, but their separate page 
> tables redirect to different real addresses.
>
>
>       [edit
>       <http://en.wikipedia.org/w/index.php?title=Virtual_memory&action=edit&section=4>]

Conclusion: not really clear definition of both terms.

My proposal:
* Logical is opposite of physical: logical always requires mapping to 
physical. Of course, logical can be mapped on some other logical thing 
that itself needs mapping to physical. More formally, this mapping is a 
general function: 1) nothing should prohibit mapping multiple logical 
names/addresses to a single physical address/thing, 2) not all physical 
entities must have a logical counterpart. Also the source and target of 
the mapping should not be restricted in terms of types: names can be 
mapped onto other names or names can be mapped onto address for examples.
* Virtual is a specialized form of logical, as the virtualization 
technology still needs to map a virtual entity onto a single 
non-virtualized (physical or isolated uniquely identifiable logical) 
entity, while guaranteeing isolation. More formally, this mapping is an 
injection of tupples of <entity, virtual instance ID> onto entities of 
the same type (thus not mapping of names into addresses for example: 
virtualization of name space, means mapping <virtual name, instance ID> 
on to unique identifiable <name>, virtualization of virtual memory (I 
would rather call this logical memory) page tables <virtual page number 
(seen by application), instance ID> on to <system wide page number>).

Kind regards,

Didier

--
Didier Colle
Ghent University - IMEC - IBBT
Department of Information Technology (INTEC)
Gaston Crommenlaan 8 bus 201, B-9050 Gent (Ledeberg)
Email: didier.colle at intec.UGent.be
MSN: didiercolle at hotmail.com
Skype: didiercolle
Tel. +32 9 331 4970
Fax. +32 9 331 4899
Mobile: +32 473 295655
WWW: www.ibcn.intec.UGent.be