[Gen-art] A *new* batch of IETF LC reviews - 2015-05-07

[Paul Kyzivat <[e-mail]>]

I am the assigned Gen-ART reviewer for this draft. For background on
Gen-ART, please see the FAQ at

<http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.

Please resolve these comments along with any other Last Call comments
you may receive.

Document: draft-ietf-opsawg-vmm-mib-02
Reviewer: Paul Kyzivat
Review Date:
IETF LC End Date: 2015-05-18
IESG Telechat date: (if known)

Summary: Ready with minor issues.

Major issues:

None.

Minor issues:

* Figure 2: A few things are fuzzy about this figure:

-- The meaning/purpose of the part above the "====", and its 
relationship to the rest of the diagram, isn't clear to me. Is it a 
legend, explaining the notation for transient vs. finite states?

-- what is the point of the 'preparing' state? There is no way in, and 
the only way out is to shutdown. (I could understand it as a starting 
state if there was a path to running.) While it is described later, in 
this figure it seems to have no purpose.

-- the 'blocked' and 'crashed' states have no way either in or out. 
Surely there must be some path into these states, and some path out (at 
least to shutdown or deleted.)

I see from the later definitions that arbitrary state transitions can be 
represented. Is Figure 2 intended to normatively constrain the state 
transitions? Or does it only provide an overview of "expected" transitions?

I don't feel I understand the intent sufficiently to suggest changes to 
remedy my confusion.

* Section 5

This says "Hypervisors *shall* implement HOST-RESOURCES-MIB." This 
sounds normative. If so, 'shall' should be replaced with MUST.

The same issue with 'shall' is present in the 2nd paragraph refering to 
virtual machines.

Also in the 2nd paragraph I can't parse or fully understand the last 
sentence. ("This document defines the objects of these information.") 
Changing 'these' to 'this' would make it grammatical, but still not very 
clear. I guess you mean something like: "This document defines the 
relationship between the objects visible to virtual machine operators 
and those visible to hypervisor operators."

* Section 8 - Security Considerations:

I see no 2119 language in this section, but I see language that sounds 
normative to me. E.g., "When SNMPv3 strong security is not used, these 
objects ***should*** have access of read-only, not read-write." If these 
statements are intended to be normative then please use 2119 language.

The rest leaves me concerned about security. But I will leave it to a 
security review to dig into.

Nits/editorial comments:

* The introduction says that this has been derived from "enterprise 
specific" MIB modules. But the examples sound more "product-specific" 
than enterprise-specific. I guess you mean modules created by the 
enterprise producing the product, so maybe it is ok, but it struck me as 
odd. (Please feel free to leave this as-is if the usage is appropriate 
in context.)

* Page 22, DESCRIPTION of vmHvSoftware:

This says "This value should not include its version, and it should be 
included in `vmHvVersion'." IIUC 'and' is the wrong word to use here - 
'as' would convey the intended meaning.