Fog Computing: A Platform for Internet of Things and Analyticspalencar/cs846/fall-2016/presentations/... · What is Cloud Computing? •Provides IT resources (storage, processing,

Fog Computing: A Platform for Internet of Things and Analytics

AUTHORS: FLAVIO BONOMI, RODOLFO MILITO, PREETHI NATARAJAN AND J IANG ZHU

PRESENTER: JOSH JUNG

FOG COMPUTING: A PLATFORM FOR INTERNET OF THINGS AND ANALYTICS 1


What is Cloud Computing?

What is Cloud Computing?•Provides IT resources (storage, processing, networking) from a centralized location


ProsNo initial investmentNo dealing with hardwareEconomy of scale

ConsPotentially high latencyContinuous costsSecurity of data


What is Fog Computing?

What is Fog Computing?•Extends the Cloud paradigm to the edge of the network

•Provides the same services in a hierarchical, distributed system




Cloud

Endpoints

Cloud Computing Fog Computing

Cloud

Endpoints

Core

Edge





What is Fog Computing?•Main advantages

–computing resources near to endpoints

–can accommodate moving endpoints

–consistent interface for heterogeneous hardware


How does this apply to Big Data?•Applies to the Internet of Things (IoT)


“The big N in these scenarios is neither the number of terrabytes nor rate of data generated by any individual sensor, but rather the number of sensors that are naturally distributed, and that has to be managed as a coherent whole.”

Use Case: Smart Traffic Light System (STLS)


•3 Goals (in order of importance):1. Accident prevention

◦ Real-time → Edge level

2. Maintenance of steady traffic flow (ie. “green wave”)◦ Near real-time → Core level

3. Collection of relevant data for analysis◦ Long term → Cloud level



•Other requirements:–Common interface for varying sensors

–Consistency in aggregation points

–Multi-tenancy



Use Case: Wind Farm


Use Case: Wind Farm•4 operating conditions:

1. Low wind speed – Do not run turbine

2. Normal wind speed – Run turbine at max speed

3. High wind speed – Run turbine, but clamp speed

4. Very high wind speed – Do not run turbine


Use Case: Wind Farm•Other requirements:

–Global control policy•Cloud level

–Subsystem optimization (eg. don’t starve other turbines)•Edge level

–Mediation between subsystems and cloud•Core level


Use Case: Wind Farm•Other requirements:

–Data Analytics• Cloud level


Nuts and Bolts


Nuts and Bolts•Physical resources – pretty much anything

–Abstraction layer hides heterogeneity and provides multi-tenancy

•Software–Foglet (IOx) runs on all nodes

•Distributed database is supported


Nuts and Bolts


•Cloud is a (potentially) hybrid collection of other cloud providers

Nuts and Bolts

•Administrators control applications and define policies via Fog Director


Nuts and Bolts•Many policy types are supported:

–Specify thresholds for load balancing

–Set Quality of Service requirements

–Manage power usage

–Specify security and privacy settings

–Configure individual devices


Nuts and Bolts


Yes Great, do it.

Fog: Pros and Cons (compared to Cloud)


Pros•Potentially Lower latency•Can accommodate moving endpoints•Better when Internet connection is unreliable•Consistent interface for heterogeneous hardware•Less data sent to Cloud

Cons•Requires some initial investment

•Have to deal with hardware

•Non-trivial amount of extra software development

•Still have to pay for Cloud usage

•Requires expansion with growing user base

References

Flavio Bonomi, Rodolfo Milito, Preethi Natarajan, and Jiang Zhu. Fog computing: A platform for internet of things and analytics. In Big Data and Internet of Things: A Roadmap for Smart Environments, pages 169-186. Springer, 2014.

Marcelo Yannuzzi, R Milito, Rene Serral-Gracia, D Montero, and Mario Nemirovsky. Key ingredients in an iot recipe: Fog computing, cloud computing, and more fog computing. In 2014 IEEE 19th International Workshop on Computer Aided Modeling and Design of Communication Links and Networks (CAMAD), pages 325-329. IEEE, 2014.

Cisco fog data services - products services. http://www.cisco.com/c/en/us/products/cloud-systems-management/fog-data-services/


Additional Image Sourceshttp://weknowyourdreams.com/cloud.html

http://www.bahrainweather.gov.bh/education_fog

http://searchengineland.com/intersection-search-social-186573

http://teachnuclear.ca/all-things-nuclear/energy-demand-and-sources/present-energy-sources/wind_energy/


http://weknowyourdreams.com/cloud.html

http://www.bahrainweather.gov.bh/education_fog

http://searchengineland.com/intersection-search-social-186573

http://teachnuclear.ca/all-things-nuclear/energy-demand-and-sources/present-energy-sources/wind_energy/

Discussion: Strengths and Weaknesses•Strengths:

–Fog seems legitimately useful in some situations–Well-motivated

•Weaknesses:–2/3 of paper spent on justification–Descriptions are incredibly vague–Use cases are purely speculative–No numerical evidence to support claims–Drawbacks of Fog are ignored


Discussion: Related PapersOriginal Fog Computing paper:

Flavio Bonomi, Rodolfo Milito, Jiang Zhu, and Sateesh Addepalli. Fog computing and its role in the internet of things. In Proceedings of the first edition of the MCC workshop on Mobile cloud computing, pages 13-16. ACM, 2012.

Follow-up paper:

Marcelo Yannuzzi, R Milito, Rene Serral-Gracia, D Montero, and Mario Nemirovsky. Key ingredients in an iot recipe: Fog computing, cloud computing, and more fog computing. In 2014 IEEE 19th International Workshop on Computer Aided Modeling and Design of Communication Links and Networks (CAMAD), pages 325-329. IEEE, 2014.

Dew Computing:

Karolj Skala, Davor Davidovic, Enis Afgan, Ivan Sovic, and Zorislav Sojat. Scalable distributed computing hierarchy: Cloud, fog and dew computing. Open Journal of Cloud Computing (OJCC), 2(1):16-24, 2015.


Discussion: Future Work•Numerical comparison of task completion time for Fog vs. Cloud

•Comparison of development time for projects using Fog vs. Cloud

•Flesh out ideas for mobile endpoint support

•Could Fog be viable as a general utility?

•Come up with more water-based names


Discussion: Additional Questions•Given that the “big” here refers to geographical distance, is it reasonable to call this Big Data?

•Can you think of other domains where Fog Computing would provide an advantage?

•Is the market for this technology too niche for it to generate much academic interest?


Fog Computing: A Platform for Internet of Things and Analyticspalencar/cs846/fall-2016/presentations/... · What is Cloud Computing? •Provides IT resources (storage, processing,

Documents