Cloud Computing for Social Inclusion
Paul Watson
Digital Institute & School of Computing Science,
Newcastle University, UK
Funders: RCUK Digital Economy Programme (SiDE), Microsoft, Red Hat, EU (Venus-C)
Social Exclusion
The result of related factors
that prevent individuals or groups
from participating fully in the
economic, social & political life
of society
• €15M Digital Economy Research “Hub”
– Funded by the UK Research Councils
– 2009-2014
• based at Newcastle & Dundee Universities
Plan
• How can Digital Technologies transform the
lives of excluded people?
• How can Cloud Computing transform Science?
Activities
Accessibility
Perspectives
Technology
Society
Business
Users
Design
Connected Transport
Home &
Community
Creative
Industries
Activities
Accessibility
Perspectives
Technology
Society
Business
Users
Design
Connected Transport
Home &
Community
Creative
Industries
Accelerometer Video
http://www.youtube.com/watch?v=hlmsrJOheS
8&list=UUiYv8UGLm8KEIM0J4GlbIw&index=5&feature=plcp
Analysis
Sleep
Clinician’s
Report
Activity
Stability
Information
for
Users
Grip
…
…
Methodology
section for
papers
Analysis
Sleep
Activity
Stability
Grip
…
…
Cloud
Clinician’s
Report
Information
for
Users
Methodology
section for
papers
Cloud Computing
• Opportunity to revolutionise IT (and Science)
– On-demand resources
– Pay-as-you–go
Cloud Computing
• Opportunity to revolutionise IT (and Science)
– On-demand resources
– Pay-as-you–go
• But Major Barriers
– Building Cloud-based systems
– Security
Cloud Computing
• Opportunity to revolutionise IT (and Science)
– On-demand resources
– Pay-as-you–go
• But Major Barriers
– Building Cloud-based systems
– Security
• Our work to address this:
Building on Cloud Infrastructure
App n
App 1
....
Cloud Infrastructure:
Storage & Compute
Problems
App n
App 1
....
Science requires apps to be:
scalable,
reliable,
secure
This requires:
expertise
Cloud Infrastructure:
Storage & Compute
time,
money,
Cloud Catch 22
Most Projects &
Organisations
that could benefit
most from
the cloud
lack the IT skills
to do so
App n
App 1
App 1
....
App n
....
Cloud Platform
Cloud Infrastructure:
Storage & Compute
Cloud Infrastructure:
Storage & Compute
App
....
App
API
Security
Workflow
Enactment
Analysis Services
Social
Networking
Provenance/
Audit
Metadata
<expt>9127</expt>
<smiles>CHC3</smiles>
Processing
Storage
Cloud Infrastructure:
Amazon, Azure,
Private Clouds
e-Science Central Video
http://www.youtube.com/watch?v=3rW2W3cL0U
applications
• UK National X-ray photoelectron spectroscopy
service
• speech to text applications
• Supporting computer games for rehabilitating
stroke victims
• machine learning
• neuroscience
Scaling Response Time:
460K workflow executions
4.4M service calls
200 Nodes
5yrs  10 hours
Cloud Security Challenges
Patient Data
Accelerometer
Data
A. Smith
378456729
Anonymize
d0
s1
d2
Results
Analyze
p = 30%
q = 27.4
r = 34
s3
d4
Public XOR Private Clouds
Application
e-Science Central
e-Science Central
e-Science Central
Amazon
Azure
Private Cloud
Problem
• Can’t exploit multiple clouds in one workflow
Patient Data
Accelerometer
Data
A. Smith
378456729
Anonymize
d0
s1
d2
Results
Analyze
p = 30%
q = 27.4
r = 34
s3
d4
Method
(P. Watson, A Multi-Level Security Model for Partitioning
Workflows over Federated Clouds IEEE CloudCom 2011)
1.
2.
3.
4.
5.
6.
7.
8.
Assign Security Level to each Workflow Block
Check conforms to Bell-LaPadula
Assign Security Level to each Cloud
Determine possible allocations of blocks to clouds
Determine candidate workflow partitioning
Add inter-cloud data transfers
Filter
Apply Cost Model to Rank candidate solutions
http://www.cs.ncl.ac.uk/publications/trs/papers/1271.pdf
Bell LaPadula for Workflows
d0
No Read Up
s1
d2
No Write Down
1. Assign Security Level to each
Workflow Block
Patient Data
A. Smith
378456729
Heart Rate Data
Anonymize
Results
Analyze
p = 30%
q = 27.4
r = 34
d0
s1
d2
s3
d4
Location: 1
0
0
0
0
Clearance:
1
0
3. Assign Security Level to each Cloud
Private
C1
Location: 1
Public
C0
0
Extend Bell-LaPadula so a block cannot be
deployed on a cloud with a lower security level
d0
s1
d2
pa
pb
pc
5. Determine candidate workflow
allocations
6. Add Inter-Cloud Transfers
6. Add Inter-Cloud Transfers
7. Filter
copy
d0
onto
pa
copy
d2
onto
pb
Valid Workflows
8. Apply Cost Model to Rank candidate
solutions
Data Costs
8. Apply Cost Model to Rank candidate
solutions
CPU Costs
8. Apply Cost Model to Rank candidate
solutions
Transfer Costs
8. Apply Cost Model to Rank candidate
solutions
8. Apply Cost Model to Rank candidate
solutions: Example 1
Workflow Costs #1
3
4
6
5
2
1
8. Apply Cost Model to Rank candidate
solutions: Example 2.
Workflow Costs #2
1
2
5
6
3
4
A Systematic Approach to Cloud Federation
Application
Security, Dependability, Performance
Requirements
Policy Manager
e-Science Central
e-Science Central
e-Science Central
Amazon
Azure
Private Cloud
How can people remain healthy and in their own
homes for longer?
Professor Patrick Olivier
Ambient Kitchen Video
http://www.youtube.com/watch?v=nRAbAlqC_o
Y&list=UUiYv8UGLm8KEIM0J4GlbIw&index=6&feature=plcp
Complex Event
Processing
Events
Activity
Recognition
Activities
POMDP…
Prompts
Prompting
• SiDE aims to realise the potential of digital technology
to make an impact on social exclusion
www.side.ac.uk
• How can we realise the potential of clouds?
– overcome barriers: application development & security
• e-Science Central aims to do this
– Ease of Use, Security, Scalability
– Exploiting Federated Clouds