CLOUD COMPUTING FOR MOBILE USERS:
CAN OFFLOADING COMPUTATION
SAVE ENERGY?
Appears in: 2010, Computer, IEEE Computer Society
Authors: Karthik Kumar and Yung-Hsiang Lu
Electrical and Computer Engineering , U.S, Purdue University
Speaker: Allen Liao
OUTLINE
INTRODUCTION
SAVING ENERGY FOR MOBILE SYSTEMS
OFFLOADING COMPUTATION TO SAVE ENERGY
ENERGY ANALYSIS FOR COMPUTATION OFFLOADING
SAMPLE APPLICATION BENEFITING FROM OFFLOADING
MAKING COMPUTATION OFFLOADING MORE ATTRACTIVE
CHALLENGES AND POSSIBLE SOLUTIONS
PRIVACY AND SECURITY
RELIABILITY
REAL-TIME DATA
CONCLUSION
2
INTRODUCTION
Cloud Computing
A service provider owns and manages resources ( such as processing,
memory, storage ), and users access them via the Internet.
For example, Amazon Web Services
Simple Storage Service (S3): let users store personal data
Elastic Compute Cloud (EC2): perform computations on stored data
3
INTRODUCTION
Advantages
Low initial capital investment
Shorter start-up time for new services
Lower maintenance and operation costs
Higher utilization through virtualization
Easier disaster recovery
There are several benefits in shifting computing from the
desktop to the cloud.
The primary constraints for mobile computing are
limited energy and wireless bandwidth.
Can cloud computing provide energy saving as a service
to mobile users?
4
SAVING ENERGY FOR MOBILE
SYSTEMS
Mobile systems
Various studies have identified longer battery lifetime as
the most desired feature of such systems.
Many applications are too computation intensive.
The computation must be performed in the cloud.
Other applications can run on a mobile system. However,
they consume significant amounts of energy.
Image retrieval, voice recognition, gaming, and navigation.
5
SAVING ENERGY FOR MOBILE
SYSTEMS
Low-power design has been an active research topic for many years.
Basic approaches to saving energy and extending battery lifetime in
mobile devices:
1.
Adopt a new generation of semiconductor technology.
As transistors become smaller, each transistor consumes less power.
More transistors are needed to provide more functionalities and better
performance
2.
Avoid wasting energy.
Whole systems or individual components may enter standby or sleep modes to
save power.
3.
Execute programs slowly.
When a processor’s clock speed doubles, the power consumption nearly octuples.
If the clock speed is reduced by half, only one quarter of the energy is consumed.
4.
Eliminate computation all together.
The mobile system does not perform the computation.
Instead, computation is performed somewhere else.
6
OFFLOADING COMPUTATION TO SAVE
ENERGY
What distinguishes cloud computing from the client-server
computing model is the adoption of virtualization.
Client-server computing: service providers managing
programs running on servers
Cloud computing: Allows cloud vendors to run arbitrary
applications from different customers on virtual machines.
Cloud vendors thus provide computing cycles, and users can
use these cycles to reduce the amounts of computation on
mobile systems and save energy.
Cloud computing can save energy for mobile users through
computation offloading
Virtualization: Lets applications from different customers run on
different virtual machines, thereby providing separation and
protection.
7
ENERGY ANALYSIS FOR COMPUTATION
OFFLOADING
Suppose the computation requires C instructions.
M: The speeds in instructions per second of the mobile system
S: The speeds in instructions per second of the cloud server
D: The bytes which the server and mobile system exchanged
B: Network bandwidth
Pc: The mobile system consumes for computing
Pi: The mobile system consumes while being idle
Ptr: The mobile system consumes for sending and receiving data.
8
ENERGY ANALYSIS FOR COMPUTATION
OFFLOADING
9
ENERGY ANALYSIS FOR COMPUTATION
OFFLOADING
Suppose the server is F times faster
S=F×M
For example
HP iPAQ PDA
M = 400 MHz, Pc ≈ 0.9 W, Pi ≈ 0.3 W, and Ptr ≈ 1.3 W.
Four-core server
S = 3.2 GHz, F = (S/M) ≈ [(3.2 × 1,024 × 4 × X)/400]
Where X is the speedup due to additional memory
If X = 5, F ≈ 160
10
ENERGY ANALYSIS FOR COMPUTATION
OFFLOADING
Bo: The minimum bandwidth required for offloading to save energy
11
SAMPLE APPLICATION BENEFITING
FROM OFFLOADING
Chess
Markovian: the game is fully expressed by the current state.
The amount of computation for chess is very large.
Can be parallelized, making the value of F very large.
The amount of computation C is extremely large, and D is very small, offloading
is beneficial for most wireless networks.
An image retrieval application retrieves images
Features: The program accomplishes this by comparing numerical
representations of the images.
The features for the image collection can be computed in advance.
The value of C is small.
D is large since considerable data must be sent.
Even if the values of F become ∞, D/B might still be too large when compared to
C/M
12
Offloading saves energy only if B is very large
At high bandwidths
MAKING COMPUTATION OFFLOADING
MORE ATTRACTIVE
Energy saved by computation offloading
B: Wireless bandwidth
C: Amount of computation to be performed
D: Amount of data to be transmitted
Fundamental assumption
Client-Server Model: Because the server does not already contain
the data, all the data must be sent to the service provider.
Cloud Computing: The cloud stores data and performs computation
on it.
Google’s Picasa
Amazon S3
Amazon EC2
There is no longer a need to send the data over the wireless network;
it suffices to send a pointer to the data.
13
A very small D and very large F imply that energy can always be saved.
CHALLENGES AND POSSIBLE
SOLUTIONS
PRIVACY AND SECURITY
Because the data is stored and managed in the cloud, security and
privacy settings depend on the IT(information technology)
management of the cloud provides.
A bug or security loophole in the cloud might result in a breach of
privacy.
For example
March 2009, a bug in Google caused documents to be shared
without the owners’ knowledge.
July 2009, a breach in Twitter allowed a hacker to obtain
confidential documents.
2007, a phishing attack duped a staff member for salesforce.com
into revealing a password.
14
PRIVACY AND SECURITY
“Tracking” of individuals through location-based
navigation data offloaded to the cloud.
Data stored at one location may not be secure at
another due to different access rules.
The EU has stricter privacy protection than the US
Some types of data cannot be stored in the cloud
without considering these privacy and security
implications.
One possible solution is to encrypt data before storage.
15
PRIVACY AND SECURITY
Pc × (Cp/M): The additional energy required to protect privacy and
security.
If this value is significant, cloud computing might not save energy.
16
RELIABILITY
A mobile user performing computation in the cloud
depends on
The wireless network
Cloud computing may not be possible, when connectivity is limited.
National parks
Basement of a building
The value of B can become very small or even zero.
Cloud service
Cloud for important computations could lead to problems during
service outages.
17
Google
Amazon
T-Mobile
Significantly reduce the value of F.
RELIABILITY
Data storage
In October 2009, T-Mobile’s and Microsoft’s mobile
Sidekick service crashed
All customers lost their data and contacts.
Independent backup of data with an alternate service
provider
Might increase the value of D
18
REAL-TIME DATA
Real-time data: D is no longer a pointer to the data; it
refers to the actual data.
Chess
Mobile surveillance
Context-aware navigation
When the value of D is large, offloading may not save
energy.
Partitioning computation between the mobile system and
the cloud to reduce energy consumption.
19
CONCLUSION
Analysis suggests that cloud computing can potentially
save energy for mobile users.
Not all applications are energy efficient when migrated to
the cloud.
Mobile cloud computing services would be significantly
different from cloud services for desktops
Offer energy savings
Before offloading, the services should consider energy
overhead for
20
Privacy
Security
Reliability
Data communication