[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [saf] SAF Use cases for CEP, GreenIT, Energy, etc
Thanks Mike. First a quick summary (to get it straight in my mind) and
then a few questions/comments. Energy Provider contributes SAF Protocols for “increase
capacity”, “reduce capacity”. Data Center (consumer) contributes SAF Syndromes to
detect “energy budget overage”. Vendor devices (in the data center) contribute Symptoms
about energy usage (so as to trigger the above Syndrome). City planners glue it all together (associate the
Syndromes to Protocols). Quetsions/comments: 1. If
the DataCenter detects an “over budget” condition, would they really ask the
energy provider to “reduce capacity”, by presumably shutting off data center ‘smart’
devices? I would think the DataCenter would want control of that type of
stuff. 2. Pursuant
to above, I’m thinking this might work the ‘other way ‘round’. Energy Provider
detects the ‘energy overage’ and generates a symptom to the DataCenter
informing them of such. The DataCenter decides (in their Syndrome) whether to
take action (Protocol) and what type of action, ie: take some devices offline. The
Syndrome would consider the EnergyProvider ‘energy overage’ symptom, along with
other internal/external symptoms such as ‘big sales opportunity’ symptom (that
might defer the action temporarily). 3. I’m
a bit ignorant about device energy usage, so forgive me. But, in the real
world, would we be so granular as to monitor the energy consumption of every
device, or would we monitor the data center energy consumption as a whole. In
other words, the Symptoms would be coarse (only at datacenter level), but
protocols would be granular (knowing which specific devices could be taken
offline, and understanding their power consumption and the effects thereof). 4. I
got a little lost in how the City planners & FEMA fit into the picture. Is
this a bigger picture where the City is facing a brownout, and the planners
must decide which DataCenters survive and which are shutdown? By the way – I think this #4 (if that’s what you
intended), is really cool scenario. SAF might use this to re-engage with the
Oasis Emergency Management TC at some point… although extended to consider not
just datacenters, but hospitals, etc. -----Original Message----- Here's my Energy/Green IT use case - appreciate feedback Abstract: As IT customers and cities look at challenges
of addressing restrictions on their overall energy utilization and
their ability to react to changing utility prices and supply constraints, the
instrumentation of the energy delivery system enables symptoms to enhance
the early detection of outages and overruns and taking preventive actions to
avoid larger outages. Making these sensitive to the business
operations and other priorities (e.g. maintaining emergency facilities) are
critical requirements so that tradeoffs can be made by both the
consumer and the suppliers. One interesting aspect to consider is that
maintenance schedules for equipment is an important element of
ensuring resources which can affect the availability of the Grid are current on
maintenance so monitoring compliance and factoring that into overall
syndrome. Use Case 1: Automation of Energy Distribution
Optimization and Outage Avoidance - Actors: Energy Utility Provider, Data
Center Customer, Vendor Suppliers, City Planner, FEMA - Policy: Energy budget for Customers,
priorities for who to reduce when energy reaches threshhold - Dynamic Composition of SAF Catalog: o Energy Provider contributes Protocols,
ie: “increase capacity”, “reduce capacity”, etc. for
businesses, homes, government facilities (e.g. nuclear power
plant) o Business (DC) Consumer contributes
Syndrome (with signature based upon Symptoms he will emit), Data
Center Over Energy Budget * Vendor supplier publishes list of
symptoms they emit (e.g. Power supply failure, Server over
Budget, Network over Budget, etc.) * Vendor supplies list of protocols
"schedule maintenance" o City Planner and FEMA gets list of
advertised Protocols from the Catalog (monitor in context of
overall GRID and special consumers). o City Planner references “reduce capacity”
Protocol to his “Avoid GRID outage” Syndrome. o Customer activates Syndrome "avoid
Data Center Outage". - Business <-> Operations Bridge: o Business elements § Protocols for “increase capacity”,
“reduce capacity” § Syndrome for “Power outage/brownout” * Syndrome for "GRID
overload" o Corresponding Operational elements § Prescription associated with
“increase capacity” · Evaluate short-term
load in context of budget · If load is predicted to
be low – deny; Otherwise… · Allocate VM from
resource pool · Bring online § Symptoms associated with
“avoid outage” · Workload performance
(track compliance to SLAs) · System performance and
utilization Michael E. Baskey IBM Distinguished Engineer Chief Architect Systems Management Standards +Open Source
Strategy TC Member IBM Academy of Technology Chairman DMTF Board of Directors 845-435-1927 or 8-295-1927 mbaskey@us.ibm.com |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]