The Massachusetts Open Cloud (MOC) initiative aims to create an open cloud infrastructure by integrating multiple partners for collaborative research in cloud computing and big data. Key goals include providing a neutral platform for technology partners, enhancing access to data and resources, and supporting Massachusetts' big data initiatives with funding proposals. The initiative seeks letters of support and collaboration to raise funds and enhance cloud capabilities for academic and industry users.

1. The Massachusetts Open Cloud
(MOC)
Orran Krieger (BU)
Peter Desnoyers (NEU), Daniel Kamalic (BU)
Credit/Collaborators:
John Goodhue(MGHPCC), Peter Desnoyers (NEU), Chris Hill (MIT), Azer Bestavros (BU), Daniel
Kamalic (BU), Jonathan Appavoo (BU), Alex Benik (Battery), Azer Bestavros (BU), John Byers (BU), David
Cohen (EMC), Chrys Lynch (Atlas), Gene Cooperman (NEU), Peter Desnoyers (NEU), Srini Devadas
(MIT), Shafi Goldwasser (MIT), Sharon Goldberg (BU), John Goodhue (MGHPCC), Michael Goroff, Jan
Mark Holzer (Red Hat), David Emory Irwin (UMass), Frans Kaashoek (MIT), Orran Krieger (BU), Jim
Kurose (UMass), Barney Maccabe (ORNL), Sam Madden (MIT), Jeff Nick (Pivotal), Paul Rad (Rackspace),
Andrei Ruckenstein (BU), Larry Rudolph (MIT), Margo Seltzer (Harvard), Prashant Shenoy (UMass), Salil
Vadham (Harvard), Daniel Wichs (NEU), Nickolai Zeldovich (MIT), Michael Zink (UMass)…

2. Cloud computing
•
Clouds having a dramatic
impact:
•
•
•
Consumer: ondemand access to
inexpensive
computational capacity,
pay for what you use
Producer: economy
of scale, automation
Like power, most
computation will move
into public clouds.

3. Problems with today's “closed”
public clouds
•
Highly prescriptive in HW, computational model, economic model; focus on
scale-out web applications
•
•
Operational/performance data limited to the single provider
Limiting research, innovation by third parties
➡
➡
•
technology companies locked out of public clouds; disconnect with
private clouds
difficult for anyone else to efficiently support/innovate Big data platforms,
No visibility/auditing of internal operations:
➡
•
•
•
Major security challenge for hosting critical datasets
Accretion of features/services into Provider offering
Monoculture increasingly dangerous
Vendor lock in by features, interfaces, and pricing model.

4. A new model is required:
an “open cloud”
•
•
•
•
•
Multiple “partners”
participate in implementing
and operating cloud
Each partner determines
how to charge for her
services
Operational data visible to
stakeholders
Domain specific
“intermediaries”:
provide customers with
simple model
enable optimization
Multi-sided marketplace
•
•
Web
HP
HP
Big Data
...
HPC
Rackspace
Red Hat
...
SeaMicro
Quanta

5. The Opportunity
•
•
•
•
•
•
OpenStack provides most of what we need:
•
modular structure with multiple independent services and
support for plugins
15 MW MGHPCC data center, low power cost, excellent
network connectivity...
MGHPCC consortium: BU, MIT, NE, UMass, Harvard.
•
•
Operate production cloud capacity for research
computation & enable Big Data & HPC users
Enable research in Big Data, Cloud Computing
Incredible regional cluster of technology companies and
innovative users of technology
Commonwealth Big Data Initiative
Launched attempt to create “Massachusetts Open Cloud
(MOC)” as a partnership: State, MGHPCC, Industry

6. Value to Technology
Partners
•
A neutral platform where private-cloud participants can
integrate, test and certify their HW and SW
•
Access to users and rich data about how products are used
•
Demonstrate technologies to be sold to private and public
clouds
•
Evaluate new products with real customers at an early stage
•
A platform to engage with the broad research community in
the participating institutions
•
Access to a community of students across the institutions
working on Cloud Computing and Big Data
•
An environment to demonstrate value to State, Federal

7. Status
• Key
part
of
State’s
big
data
ini4a4ve:
LOI
for
$3M:
will
host
the
states
public
data
sets
and
enable
startups
in
Big
Data...
• MLSC
funding
$4.5M
to
create
cloud
for
life
sciences
users
to
advantage
research
and
local
life
science
industry
• Approval
from
consor4um
to
use
HGHPCC
15MW
data
center
• ORNL
puPng
together
plan
to
par4cipate.
• By
the
end
of
October
full
proposal
to
State
for
$3M,
need
to
raise
at
least
$9M
in
matching
involvement
from
founding
partners:
– Pending
&
exis4ng
commits:
XXXX
– In
conversa4on:
XXXX
– >
$6M
locked
down,
expect
remaining
commitments
to
come
in
on
4me
• In
a
good
posi4on
to
compete
for
$10M
NSFcloud
testbed
grant
in
Dec
7

8. Our
ask
• Leers
of
support
(by
next
week)
• Use
cases
• Development
collabora4on
• People
interested
in
posi4ons...
8