Presentation to CleanTech Future Conference II in San Francisco, 4 November 2013, on multi-tenancy's 95% reduction of IT CO2 footprint - versus timid incrementalism of virtual-machine approach
1. Redefining “Clean IT”
Rejecting Incremental Improvement
Peter Coffee
VP / Head of Platform Research
salesforce.com inc.
CleanTech Future Conference II
San Francisco, November 2013
2. ICT Wattage: the Globe-Warming Numbers
• Global information/communication technology (ICT) “uses
1,500 terawatt-hours of power per year. That’s about 10% of
the world’s total electricity generation or roughly the
combined power production of Germany and Japan.”
- science.time.com/2013/08/14/power-drain-the-digital-cloud-is-using-more-energy-than-you-think
• “Data centers can waste 90% or more of the electricity they
pull off the grid… they further rely on banks of generators…
In Silicon Valley, many data centers appear on the state
government’s Toxic Air Contaminant Inventory, a roster of
the area’s top stationary diesel polluters.”
- www.nytimes.com/2012/09/23/technology/data-centers-waste-vast-amounts-of-energy-belying-industry-image.html
3. Measuring the Wrong Thing…Precisely
• “Power usage effectiveness (PUE) is a measure of how
efficiently a computer data center uses energy;
specifically, how much energy is used by the computing
equipment (in contrast to cooling and other overhead).”
- en.wikipedia.org/wiki/Power_usage_effectiveness
• PUE is easy to determine, but is absolutely the wrong
way to define ICT ‘greenness’
• Treats every watt-hour burned by computers as equal-value
• Ignores utter redundancy of running multiple, identical instances
of operating system; database engine; & other foundationware
4. Peter Drucker (as usual) Nails It:
• “There is nothing so
useless as doing
efficiently that which
should not be done at all.”
• “Efficiency is doing the
thing right. Effectiveness
is doing the right thing.”
• “What's measured, improves”
- www.goodreads.com/author/quotes/12008.Peter_F_Drucker
5. So, What Should We Do? And Measure?
• Putting thousands of software
stacks in one big place, and
calling it a cloud, is like putting
wings on the QE2 and calling it an A380
• Don’t preserve complexity, and compound with more
complexity, to minimize adaptation to new environment
• Massive economies arise from massive scale:
Not this…
but this:
6. Multi-Tenancy Transforms Sustainability
Legacy IT Model
(on-prem or off)
App 1
App Server
Database
OS
Server
Storage
Network
App 2
App Server
Database
OS
Server
Storage
Network
Multi-Tenant Cloud
Our infrastructure
Other apps
App 3
App Server
Database
OS
Server
Storage
Network
• Each app has its own dedicated software stack
• Each stack needs duplicative maintenance
• Every app and its stack are individual
opportunities for error or misconfiguration
• Pace of innovation is slowed by need for
cumbersome regression test / re-implementation
• Applications share an “instance” of the platform
• Platform updates apply to all at once
• Massive economies of talent,
energy and cost
7. Architecture for Ecosystems:
Designed to Connect
User Interface
Your Clicks
Logic
Your Code
Salesforce to
Salesforce Sharing
Database
Selectively shared data, logic and customizations
Coherent Code Base and Managed Infrastructure
8. Multi-Tenant Cloud Computing:
Most Sustainable IT Model in the World
Energy Efficiency Comparison:
CO2/Transaction, not /Cycle or /Server
On-Premise
‘Private Cloud’
Carbon Footprint
(g. CO2 / transaction)
95%
lower
carbon
intensity
64%
lower
carbon
intensity
Estimated avoided carbon emissions from salesforce.com customers running applications on the multi-tenant
cloud as opposed to running on-premise servers. Actual carbon emissions savings could vary. Based on WSP
comparison model and research commissioned by salesforce.com, March 2011.