Mobile Cloud Computing
T-109.4300 Network Services Business Models 23.03.2011
Yrjö Raivio
Aalto University, School of Science
Department of Computer Science and Engineering Data Communications Software
Email: yrjo.raivio@aalto.fi
19.1 Introduction, Sakari Luukkainen 26.1 Mobile market, Sakari Luukkainen
2.2 Theoretical frameworks, Sakari Luukkainen
9.2 Business model design using STOF, Sakari Luukkainen & Antero Juntunen 16.2 Open Telco, Vesa Suikkola
Week 8 Winter holiday
2.3 Mobile services research, Antero Juntunen Week 10 Exam week
16.3 Cloud computing, Sakari Luukkainen 23.3 Mobile cloud computing, Yrjö Raivio 30.3 Green computing, Teemu Muukkonen 6.4 Online music business, Heikki Kokkinen 13.4 Google business model, Matti Leppänen 20.4 API brokering, Alberto Vila Tena
25.5 Examination
Course agenda
• Research motivation
• Architecture
• Hybrid Cloud
• Telecom Cloud
• Mobile Offloading
• Open Telco
• Summary
Agenda
Research motivation
Cloud markets drivers
Cloud industry opportunity (near term)
Traffic load varies also in telecom applications
Source: P. Zerfos, X. Meng, S. H.Y. Wong , V. Samantaand S. Lu, “A study of the Short Message Service of a nationwide cellular network”, IMC 2006.
Private cloud
Public cloud
Service Level Agreement – Telecom vs. Cloud
I Support systems II Tactical systems
III Strategic
systems
Ca rr ie r Gr ade
SLA Carrier grade 6 EC2 Large VMs
Availability 99.999 % 99.95 % one zone
99.9999 % two zones
Latency < 150 ms < 50 ms (EU zone)
Throughput > 1000 messages/s >1000 messages/s
Source: M. Murphy, ”Telco Clouds” [presentation], Cloud Asia 2010
Source: R. Paivarinta and Y. Raivio, ”Performance Evaluation of NoSQL Cloud Database in a Telecom Environment”, Closer 2011.
Mobile capabilities
Source: Kemp et al., ”Cuckoo: a Computation Offloading Framework for Smartphones”, MobiCASE 2010.
Bottleneck:
Battery
Architecture
Mobile Cloud
Public cloud Private
cloud
Telecom Cloud SaaS PaaS IaaS
Support Systems (OSS/BSS) Service
Delivery Storage Computation Communication
Open Telco SaaS
PaaS IaaS SaaS
PaaS IaaS
Hybrid Cloud
Mobile Offloading
End
users
• Open Source SW IaaS alternatives
• Eucalyptus (Eucalyptus Systems)
• Open Stack (Nasa, Rackspace, Cisco..)
• Open Nebula (C12G Labs)
• Interoperability of clouds: DeltaCloud
Infrastructure as a Service (IaaS)
Hybrid Cloud
Private, Public and Hybrid Cloud vs. product cycles
Time Turnover
Private cloud
Public cloud
Hybrid cloud Early
development Exponential
growth
Mature
market
Business optimization of
Dynamic Hybrid Cloud – case SMSC
• Compare
• Public cloud (Amazon EC2)
• Own private cloud (OpenNebula)
• Hosted private cloud (OpenNebula run from Web hotel)
• Dynamic Hybrid cloud (Amazon EC2
& own or hosted OpenNebula)
• Minimize
b a
F = ∫ f(y) + B ∫ f(y) dt, where 0 b
A = private cloud cost/msg B = public cloud cost/msg a = peak load
b = private cloud max capacity
• www.cloudonomics.org
Private cloud Public cloud
a
b
F (cost)
Dynamic Hybrid Cloud – case SMSC
• Basic idea: use private clouds for base load, public clouds for peaks
• Scalability: Auto scaling feature can be used to start SMSC’s on demand
• Hybrid architecture helps to achieve SLA requirements and improves trust
• Case SMSC Finland
• 3.5 billion SMS messages/year
• Base load 100 messages/s
• Peak: x10 (Tickets: x1000)
• Costs with Amazon EC2 (6 x Large virtual machine incl. transmission and storage costs): ~10 k€/year
• SMS revenues: ~200 M€/year
SMS users
Public Cloud Amazon EC2 SMSC Cluster Open-
Nebula Front
End
Private Cloud
OpenNebula
SMSC Cluster
Telecom Cloud
HLR in Cloud
HLR MSC
Client N MSC Client 2
MSC Client 1
…
Source: R. Paivarinta and Y. Raivio, ”Performance Evaluation of NoSQL Cloud Database in a Telecom Environment”, Closer 2011.
Mobile Virtual Network Operator in a cloud
True MVNO
Weak MVNO
Reseller MNO
Cloud MVNO
Radio Access &
Network
Switching
& Network Element
Inventory &
Resource Management
Services &
Content
Billing &
Customer Care
Marketing
& Sales
Source: Kiiski and Hämmäinen, ”Mobile Virtual Network Operator Strategies: Case Finland”, ITS 2004.
Mobile Offloading
Mobile Offloading:
computation vs. communication
Source: B.-G. Chun and P. Maniatis, ”Augmented Smartphone Applications Through Clone Cloud Execution”, HotOS 2009.
Source: K. Kumar and Y.-H. Lu, ”Cloud Computing for Mobile Users: Can Offloading Computation Save Energy ”, Computer, April 2010.
Mobile Offloading alternatives
Source: B.-G. Chun and P. Maniatis: ”Augmented Smartphone Applications Through Clone Cloud Execution”, HotOS 2009.
Mobile
Cloud
Model-View-Controller approach
• Extension of Model-View- Controller (MVC) pattern,
supported by Qt and other major application frameworks
• Replicate the View and part of the Controller on a User Agent
residing on another device
• Keep the User Agent and the
application synchronized using an event-propagation mechanism
Source: V. Stirbu, ”A RESTful Architecture for Adaptive and Multi-device Application Sharing”, WS-REST 2010.
Mobile Offloading use cases
A) Distributed Presenter
B) Dropbox
C) Cloud Assisted Gaming
Open Telco
Mobile network as N-sided platform – case Event Experience
… …
Subsc ribe rs
Co ntent Providers
Subscription fee
Subsidization Service fee
Commission
Operators
Messaging Location Payment Click-to-
call Call notification Data conn.
profile In-app billing Remaining
credits
GSMA OneAPI Flow festival
Helsinki City Upcode
Band merchandise Tiketti
HSL
Google Maps
Innovation Wheel
Cloud
Usage Scenarios
User acceptance (VTT&Aalto)
Business Scenarios
Business acceptance (VTT&Aalto Service
Factory)
Open Telco
Technologies&Business (TeliaSonera&Aalto)
Business Development
(Aalto Venture Garage)
Software Factory
Implementation (University of Helsinki,
Aalto)
Lessons learnt APIs
Validated service ideas
Business cases Field trials
Service prototypes
User Experience
User evaluation (VTT&Aalto) Spinoffs
