Re: [nmrg] Autonomic Use Cases

[Joe Marcus Clarke <[e-mail]>]

On 3/7/14, 6:47 PM, Michael Behringer (mbehring) wrote:
> NMRG,
>
> During the meeting we mentioned the need to document use cases. This section in the definitions draft is so far empty. Since we should have that section BEFORE working out the use cases, I drafted something up here.
>
> I also realised that while we haven't really written down in the draft that the key point of this work really is to work out common infrastructure requirements. So I'm also suggesting an additional short section in the Design Goals section:
>
>        <section title="Common Autonomic Networking Infrastructure">
> 	<t><xref target="I-D.irtf-nmrg-an-gap-analysis"/> points out that there are already a number of fully or partially autonomic functions available today. However, they are largely independent, and each has its own methods and protocols to communicate, discover, define and distribute policy, etc. </t>
> 	<t>The goal of the work on autonomic networking in the IETF is therefore not just to create autonomic functions, but to define a common infrastructure that autonomic functions can use. This autonomic networking infrastructure may contain common control and management functions such as messaging, service discovery, negotiation, intent distribution, etc. A common approach to define and manage intent is also required. </t>

Is this IETF or IRTF?  I know the goal is to eventually move into the 
IETF, but what do you intend here?

> 	<t>Refer to the reference model below: All the components around the "autonomic service agents" should be common components, such that the autonomic service agents do not have to replicate common tasks individually. </t>
>        </section>
>
> Comments? Does this capture the idea well?

I think so.  I like the idea of moving beyond atomic autonomic functions 
into more of an autonomic system.  One key aspect of this system is 
self-monitoring and diagnostics (think regulation).  The idea of probes, 
sentinels, etc. play into this, and I think it should be called out in 
the list.

>
> And then, the use case section could look like this:
>
>        <section title="Guidelines for Case Studies">
> 	<t>Case studies and problem statements are mandatory to understand common requirements for autonomic functions. This section explains how case studies should be outlined and what they should describe:
>          <list style="symbols">
> 	  <t>Title</t>
> 	  <t>Problem Statement: An explanation which problem is being addressed, with information about existing solutions and their shortcomings.</t>

Nit: OF which problem is being...

> 	  <t>Intended user / administrator experience: The goal of autonomic networking is to simplify network administration and usage. Use cases should point out how their experience differs from current solutions. If a use case depends on configuration, it may include configuration samples, although obviously the goal is to reduce or eliminate configuration. </t>
> 	  <t>Intent: Strictly speaking intent is part of the administrator experience, but should probably explained explicitly with a high-level view on how the autonomic function could be defined in intent (if required). </t>

Nit: probably BE explained...

> 	  <t>Local knowledge: What the function needs to know about the capabilities of the node itself, and which local resources need to be accessed.</t>
> 	  <t>Communication requirements: The requirements for message exchange, discovery, negotiation, etc with other autonomic nodes. </t>
>          </list>
> 	</t>
> 	<t>Use cases are not required to outline a solution in detail, nor to specify precise protocol or intent details. They are used at this point to determine a consolidated approach to developping an autonomic networking infrastructure. </t>

Nit: DEVELOPING (typo)

>        </section>
>
> Comments?

I think this is a good set of high-level guidelines.  Is there a need 
for remote knowledge as well?  That is, what if a node needs to know 
topology to make the right choice?

Joe