Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen...

10
Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1

Transcript of Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen...

Page 1: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

1

Capita Selecta Distributed Systems

Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen

IBBT-DistriNet, KULeuvenSeptember 2011

Page 2: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

September 29, 2010Capita Selecta - Distributed Sysems: general

information 2

Organization & Evaluation

• First half delivered in Semester 1– From mid October to December

• Second half delivered in Semester 2– From mid February to April

• Ongoing evaluation through exercises, class interaction and presentation

• Evaluation typically executed during and immediately after the related course activities

Page 3: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

September 29, 2010Capita Selecta - Distributed Sysems: general

information 3

Course Activities

• Tutorial style introduction to key topics

• Study and discussion of research papers

• Hands-on experience with technologies

• Critical assessment of current practices and the state-of-the-art

Page 4: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

4

Module 1: Perspectives onThe Internet of Things (IoT)

Danny Hughes, Wouter Joosen, Sam Michiels

October-December 2011

Page 5: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

Wireless Sensor Networks (WSN)

• Tiny computers with low-power radios self organize to provide distributed sensing of the physical environment:– A key technology for the Internet of Things (IoT)– A growing marketplace: to top $38bn by 2014– Key apps: medical, logistics, energy conservation, military

• Driven by technology advances. Moore’s Law predicts:– A chip of the same size doubles in power every two years– Or, the size of a chip of the same power halves every two

years (i.e. our devices are getting smaller)

Page 6: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

Wireless Sensor Networks (WSN)

• WSN Characteristics:– Large scale: 1000s of nodes over wide areas– Limited Resources: 10MHz CPU, 10K RAM, 48KB flash– Dynamism: node failure, network failure, mobility

• Two key questions:– How do we implement existing protocols and middleware in such as

small space?– What new abstractions, tools and support do we need to build

applications for this extreme environment?

• A rapidly evolving subject area, which requires knowledge of many subjects you have already studied

Page 7: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

The WSN Module

• You will revisit key topics such as:– Networking– Operating Systems– Programming Models… applied in a new and challenging IOT context.

• Using these technologies you will learn to design networked embedded applications for various applications

Page 8: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

8

Module 2: Perspectives on Cloud Computing

Wouter Joosen, Eddy TruyenIBBT-DistriNet, KULeuven

February-April 2012

Page 9: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

An Architecture Paradigm Shift...

• Cloud computing as the next distributed architecture– Do Scientific and enterprise computing have

evolved to a common architecture … ?Mainframe 1970s Centralized bulk data processing,

multi-user access on terminals (keyboard-screen)

Client-server 1980s Fat clients on workstations, centralized database server

Web 1990s Multi-tier: web tier, application servers, database.com your business

SOA 2000s Expose automated business process as a web serviceB2B communication

Cloud 2010+ Centralized provision of dynamically scalable, multi-tenant software as a service over the InternetBuilds further upon SOA

Page 10: Capita Selecta Distributed Systems Danny Hughes, Wouter Joosen, Sam Michiels, Eddy Truyen IBBT-DistriNet, KULeuven September 2011 1.

An Architecture Paradigm Shift...

• Cloud computing as the next distributed architecture– Do Scientific and enterprise computing have

evolved to a common architecture … ?Mainframe 1970s Centralized bulk data processing,

multi-user access on terminals (keyboard-screen)

Client-server 1980s Fat clients on workstations, centralized database server

Web 1990s Multi-tier: web tier, application servers, database.com your business

SOA 2000s Expose automated business process as a web serviceB2B communication

Cloud 2010+ Centralized provision of dynamically scalable, multi-tenant software as a service over the InternetBuilds further upon SOA

Now they want to outsource the problem

And hopefully find a solution!

Let’s evaluate and improve this approach….

Increased decoupling: a trivial

idea, yet hard to turn into reality