|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: IPS Draft Charter updateBlack_David@emc.com wrote: > > Mark > > Thanks for the suggestion: > > > In addition the WG may consider emerging standards for storage quality > > of service and how they may be used in IP based storage. This would > include > > the specification of performance and availability service levels for block > > storage devices and the management of policies related to those service > > levels. > > Would those be IETF standards, or some other standards > body/bodies? If the latter, which ones? As you know, I am a co-chair for the Storage Network Industry Association (SNIA) Policy working group. For the benefit of others on this list, we are working on storage policies and storage service level standards. The consensus to date has been that these policies would leverage the work of the IETF Policy Framework working group as well as the Distributed Management Task Force's (DMTF's) Common Information Model (CIM) Policy Model. The standards we come up with would be incorporated into DMTF's CIM Model, but could also be standardized as a LDAP schema or model in an appropriate IETF WG. There is already a precedent for this in the current DMTF/IETF service level/policy working groups. The main IETF WG to date that has done anything relating to storage is the IP over Fibre Channel WG and it might be appropriate to standardize it there by extending its charter. However, this working group might be a more appropriate place to do this standardization as it appeals more to storage and system vendors rather than Fibre Channel switch vendors. The Policy Framework working group has already pushed policy models specific to a given technology (DiffServ, MPLS, Security) out to the respective working groups to standardize, and this could be an example of that. As far as this working group is concerned, the standards might take the form of signaled service level objectives (requesting a certain performance goal, for example) to the storage device or perhaps a protocol for negotiation of the service level. Service level metrics, such as performance, availability and perhaps even cost might be exposed through a storage MIB. Discovery could find storage devices that would provide certain service levels and even sort them by cost or distance (latency performance metric). These all involve some form of IP protocol, however, they go beyond the initial work of simple connectivity and data transfer. My intent is to allow this sort of work in our charter so that teams that are interested in this work would have a framework within which to collaborate in the IETF. -- mark
Home Last updated: Tue Sep 04 01:08:07 2001 6315 messages in chronological order |