Re: [Detnet] [Nfvrg] [5gangip] Network Slicing - a suggestion that we meet to discuss in Seoul (INTERNAL)

[<[e-mail]>]

IMO the discussion below is around resource slicing where resources are compute/connectivity/storage.

What 3GPP talks is more about network slicing which  is more of a "logical network composed by interconnection of set of network functions". Off course network slices cannot run in the air so they need resources slices and this is how I see the mapping.

FYI:

3GPP SA2 defines the following (TR 23.799)


Network Slice Template (NST):  is a logical representation of the Network Function(s) and corresponding resource requirements necessary to provide the required telecommunication services and network capabilities.
Network Slice Instance (NSI): is an instance created from a Network Slice Template (NST).
Network Slice: is a concept describing a system behaviour which is implemented via Network Slice Instance(s).



BR,

Kashif



-----Original Message-----
From: Nfvrg [mailto:[e-mail]] On Behalf Of Jeff Tantsura
Sent: 1. november 2016 22:26
To: Stewart Bryant
Cc: Dongjie (Jimmy); Mach Chen; Joe Touch; [e-mail]; [e-mail]; [e-mail]; [e-mail]; [e-mail]; [e-mail]; [e-mail]
Subject: Re: [Nfvrg] [Detnet] [5gangip] Network Slicing - a suggestion that we meet to discuss in Seoul



Network slicing has become de facto standard terminology in context of 5G.



It would be rather useful to:

-agree on terminology

-discuss data modeling, interrelated technologies - radio, packet core, transport, optical to mention some and how they could work together/form in a "slice"

-bring together experts in these, rather disjointed domains.

-many other things



Regards,

Jeff



> On Nov 1, 2016, at 12:39, Stewart Bryant <[e-mail]<mailto:[e-mail]>> wrote:

>

>

>

>> On 01/11/2016 16:55, Joe Touch wrote:

>> Hi, all,

>>

>>

>>> On 11/1/2016 8:21 AM, Stewart Bryant wrote:

>>> We were trying to pull together a problem statement for network

>>> slicing in a 5G context to understand how well the current IETF

>>> protocols address this problem, what their short comings might be,

>>> and what IETF work is necessary to have a deployable protocol suite

>>> to address this need.

>> I've seen this term around for a while, and continue to consider it

>> nonsensical.

>>

>> It originally referred to a way of cutting OS resources and tying

>> them to a distributed system in PlanetLab, which had no support for

>> network virtualization. It has since been used to refer to E2E

>> virtual nets (distinct from VPNs, which often tether individual hosts

>> to a private net over the public Internet), but all networks have

>> always had that capability.

>>

>> IMO, this term isn't related to the Internet architecture at all. It

>> needs to correctly tie the existing network overlays that the

>> Internet has supported for a very long time to OS resources.

>>

>> I.e., most of the work in defining "network slicing" needs to happen

>> inside the OS, which is out of scope for the IETF. A properly defined

>> E2E virtual network terminates at (virtual) network interfaces, which

>> is compatible with either a sliced OS or a non-sliced OS.

>>

>> Joe

>>

>

> Joe, we possibly have different ideas as to what network slicing is.

>

> My assumption is that it is a way that we can create multiple

> virtually networks over a common base network with a higher degree of

> isolation between VPNs and much higher performance guarantees than we currently achieve.

>

> If we need a new name, we need a new name, but the concept does not

> seem to align with the definition you hints at.

>

> - Stewart

>

>

> _______________________________________________

> detnet mailing list

> [e-mail]<mailto:[e-mail]>

> https://www.ietf.org/mailman/listinfo/detnet



_______________________________________________

Nfvrg mailing list

[e-mail]<mailto:[e-mail]>

https://www.irtf.org/mailman/listinfo/nfvrg