Graduate Presentation
under the esteemed guidance of
Dr.Qi Tian
Dr.Xiaoyin wang
Dr. Ali Saman Tosun
Presented by:
Sai Gandham
Papers to be covered:
1. A view of Cloud Computing
2. Network Virtualization and Software Defined Networking for Cloud Computing
3. Network Virtualization and Resource Description in Software Defined Wireless
Networks
A view of cloud computing
By
Michael Armbrust, Armando fox, Rean
Griffith, Anthony D.Joseph, Randy katz, Andy
konwinski,Gunho lee, David Patterson, Ariel
rabkin, Ion stoica, and matel zaharia
A View of Cloud Computing
 Cloud computing refers to both the applications delivered as services over the
internet and the hardware and systems software in the data centers that
provide those services.
 Services includes 1) Software as a service
2) Infrastructure as a service
3) Platform as a service
Public and Private Cloud
 Public cloud : when a cloud is made available in a “pay as you go” manner to
the general public, we call it as public cloud.
Services being sold is utility computing
 Private cloud: When a cloud is made available only to internal data centers of
a business or other organizations , not made available to
general public.
Aspects new in cloud computing
 The appearance of infinite computing resources available on demand.
 The elimination of an up-front commitment by cloud users.
 The ability to pay for use of computing resource on a short term basis as
needed and release them as needed.
Cloud computing economics
 Three particularly compelling use cases that favor cloud computing
 1) Demand of service varies with time.
 2) When the demand is unknown in advance
 3) Organizations that perform batch analytics can use the “cost associativity”
of cloud computing to finish computations faster.
Overprovisioning and underprovisioning
Comparing public clouds and private
data centers
Obstacles and opportunities for cloud
computing
Outages in AWS, AppEngine, and gmail
service
Questions ??
Network Virtualization and Software
Defined Networking for cloud computing
By
Raj jain and subharthi paul
Network Virtualization and software
defined networking for cloud computing
 Why Virtualize ?
 Virtualization in computing
 Network Virtualization
 Software defined networking
 Open application delivery using SDN
Why virtualize ?
 Sharing
 Isolation
 Aggregation
 Dynamics
 Ease of management
Virtualization in Computing
 Virtualization is not a new concept
 Computer networking is the plumbing of computing
 VLANs allow multiple departments of company to share a physical LAN with
isolation. Similarly VPN
 Significant renewed interest in network virtualization fueled primarily by
cloud computing.
Network Virtualization
 A computer network starts with a network interaction card in the host.
 NIC  layer 2(L2) network Interconnected(via bridge)  layer3 (L3)
 Internet
 Each of these components needs to be virtualized
 Multiple standards to virtualize of several of these components.
Virtualization of NICs
 For multiple Vms on the system, each VM needs its own Virtual NIC.
 One way to solve above problem is by using hypervisor software, that provides
virtual CPU and also implements as many virtual NICs as VMs.
 vNICs  vSWITCH  pNIC  pSwitch
 VNIC implementation has different standard approaches.
Virtualization of switches
 Ethernet switch has 32 – 128 ports
 Number of physical machines that needs
to be connected on L2 network is larger
 Several layers of switches need
to be used to form L2 network
Virtual LANs in CLOUDS
 Problem in cloud: Multiple VMs in a single physical machine may belong to
different clients and need to be in different VLAN
 Each VLAN span several data centers
 Solution : VXLAN, NVGRE, STT
Software Defined Networking
 SDN is latest revolution in networking innovations.
 SDN consists of four innovations
1) Separation of control plane and data plane
2) Centralization of control plane
3) Programmability of control plane
4) Standardization of application programming interface(APIs)
Separation of control and data planes
 Networking protocols are arranged in different planes : data, control and
management
 Data plane : consists of all messages that are generated by users.
 Control plane : Deals with transport of all the messages in data plane.
Generates routing tables by using different routing protocols .
 Management Plane: Keeps track of traffic statistics and states of various
network equipment.
 Key innovation of SDN is separation of Control and data plane.
 Control logic is separated and implemented in controller that prepares
forwarding table
 This reduces the complexity and cost of the switches significantly.
Centralization of control plane
 Centralization was considered bad thing until few years ago
 Now it considered as good for good reason
 Centralization of control makes sensing the state and adjusting the control
dynamically based on state changes much faster than distributed protocols
 Standby controllers can be used to take over in case of failures of the main
controller
Programmable control plane
 It is easy for the network manager to implement control changes by simply
changing the control program.
 The programmable control plane is the most important aspect of SDN
 Programmable control plane allows the network to be divided in to several
virtual networks with different policies and yet resides on shared hardware.
Standardized APIs
 SDN consists of centralized control plane with
 Southbound API for communication with hardware infrastructure
 Northbound API is for communication with network applications
SDN impact and future
 SDN is expected to make network programmable and easily partitionable and
virtualizable
 These features are required for cloud computing where network
infrastructure is shared with number of competing entities
 SDN is expected to reduce both capital expenditure and operational
expenditure
 Network of tomorrow is more programmable than today
Open Application delivery using SDN
 Current SDN based efforts are restricted to L3 and below network traffic
 It may be expanded to L3 and above layer network traffic management
 Application traffic management involves application deployment and delivery
policies.
 Application service is replicated over multiple hosts and may be partitioned
for improved performance.
Problem statement
 Most applications need to serve global audience
 Needs servers all over the world
 Cloud services provides multiple computing and storage facilities
 Problem is routing using ASP’s policies in a very dynamic multi cloud
environment is not possible
 Since ISP’s offer no service to dynamically route messages
Open Application delivery
Questions ??
Network Virtualization and resource
description in software defined wireless
Networks
By
Qianru zhou, Cheng-Xiang wang, Stephen
Mclaughlin, and xiaotian zhou
Network virtualization and resource
description in software defined wireless
networks
 Challenges in wireless networks
 Overview of existing SDWN architectures
 Network description based on RDF
 SDWN Architecture with resource description and ontologies
Challenges in wireless networks
 SDN virtualizes the network architecture and isolates data control traffic.
 The design of wireless network architecture challenging
 Must deal with physical restrictions caused by fast changing nature of wireless
channels
 Server virtualization of wireless networks is also more challenging as it has to
satisfy the requirements of both coherence and hardware isolation
SDWN Virtualization Architecture
 SDWN is about making decisions on how a connection or flow is transmitted
across network
 SDWN is to split data and control plane
 Most widely used protocol is open flow
 It configure network elements
 Provides open protocol to program the flow table in different switches and
routers
Architecture designs of SDWN
 Current SDWN research focuses on network architecture
 Existing designs often focus on different positions
 Route flow focuses on IP routing services
 FlowVisor and FlowN concentrates on slicing the network physical
infrastructure .
 OpenRoads was proposed with the intention to replace present WIFI networks
SDWN control strategies
Information model
 In SDWN, information model is the fundamental element
 Information model describes all resources of network
 This information model is foundation of network virtualization
 It describes both physical layer infrastructure and visualization.
 Information model should be Technology independent, reusable, easily
extensible and linkable to other existing model.
Semantic technology
 Semantic web all the information and services can be understood and used
both by humans and computers.
 Semantic web is composed of three elements
 Metadata, RDF, and ontology
 Metadata is the data about data
 RDF is a standard about making statements about resources
 Ontology is also known as vocabulary, describes set of classes and relationship
between classes
 In network description ontology describes a set of nodes and relation between
them.
Network semantic ontology applications
 Until now network semantic ontology languages proposed are numberless
 These languages have different grammar, different parameters and different
specificities of application
 A universally accepted language that describes the resources of SDN is not has
ben proposed
 Due to complicated and variable wireless channels environment and emerging
new technologies building ontology for wireless networks is arduous task
Performance evaluation of ontology
 The ontology evaluation is the process to determine which resources the
ontology defines correctly/incorrectly and those it does not define.
 The criteria for performance evaluation are :
1) Consistency
2) Completeness
3) Conciseness
4)Expandability
5) Sensitiveness
SDWN architecture with resource
description function module
Resource ontology
 With this ontology we can express resources in RDF triples
 “BaseStation A” “has Antenna” “Antenna 1”
Qos ontology
 We can describe Qos resources in RDF
 “Node A” “hasMSTR” “10Mb/s”
Questions ??
Thank you
References
 Michael Armbrust, Armando Fox, Rean Griffith, Anthony D. Joseph, Randy
Katz, Andy Konwinski, Gunho Lee, David Patterson, Ariel Rabkin, Ion Stoica,
and Matei Zaharia. 2010. A view of cloud computing. Commun. ACM 53, 4
(April 2010), 50-58. DOI=http://dx.doi.org/10.1145/1721654.1721672
 R. Jain and S. Paul, "Network virtualization and software defined networking
for cloud computing: a survey," in IEEE Communications Magazine, vol. 51, no.
11, pp. 24-31, November 2013.doi: 10.1109/MCOM.2013.6658648
 Q. Zhou, C. X. Wang, S. McLaughlin and X. Zhou, "Network virtualization and
resource description in software-defined wireless networks," in IEEE
Communications Magazine, vol. 53, no. 11, pp. 110-117, November 2015.
doi: 10.1109/MCOM.2015.7321979

Cloud computing and Software defined networking

  • 1.
    Graduate Presentation under theesteemed guidance of Dr.Qi Tian Dr.Xiaoyin wang Dr. Ali Saman Tosun Presented by: Sai Gandham
  • 2.
    Papers to becovered: 1. A view of Cloud Computing 2. Network Virtualization and Software Defined Networking for Cloud Computing 3. Network Virtualization and Resource Description in Software Defined Wireless Networks
  • 3.
    A view ofcloud computing By Michael Armbrust, Armando fox, Rean Griffith, Anthony D.Joseph, Randy katz, Andy konwinski,Gunho lee, David Patterson, Ariel rabkin, Ion stoica, and matel zaharia
  • 4.
    A View ofCloud Computing  Cloud computing refers to both the applications delivered as services over the internet and the hardware and systems software in the data centers that provide those services.  Services includes 1) Software as a service 2) Infrastructure as a service 3) Platform as a service
  • 5.
    Public and PrivateCloud  Public cloud : when a cloud is made available in a “pay as you go” manner to the general public, we call it as public cloud. Services being sold is utility computing  Private cloud: When a cloud is made available only to internal data centers of a business or other organizations , not made available to general public.
  • 6.
    Aspects new incloud computing  The appearance of infinite computing resources available on demand.  The elimination of an up-front commitment by cloud users.  The ability to pay for use of computing resource on a short term basis as needed and release them as needed.
  • 7.
    Cloud computing economics Three particularly compelling use cases that favor cloud computing  1) Demand of service varies with time.  2) When the demand is unknown in advance  3) Organizations that perform batch analytics can use the “cost associativity” of cloud computing to finish computations faster.
  • 8.
  • 9.
    Comparing public cloudsand private data centers
  • 10.
    Obstacles and opportunitiesfor cloud computing
  • 11.
    Outages in AWS,AppEngine, and gmail service
  • 12.
  • 13.
    Network Virtualization andSoftware Defined Networking for cloud computing By Raj jain and subharthi paul
  • 14.
    Network Virtualization andsoftware defined networking for cloud computing  Why Virtualize ?  Virtualization in computing  Network Virtualization  Software defined networking  Open application delivery using SDN
  • 15.
    Why virtualize ? Sharing  Isolation  Aggregation  Dynamics  Ease of management
  • 16.
    Virtualization in Computing Virtualization is not a new concept  Computer networking is the plumbing of computing  VLANs allow multiple departments of company to share a physical LAN with isolation. Similarly VPN  Significant renewed interest in network virtualization fueled primarily by cloud computing.
  • 17.
    Network Virtualization  Acomputer network starts with a network interaction card in the host.  NIC  layer 2(L2) network Interconnected(via bridge)  layer3 (L3)  Internet  Each of these components needs to be virtualized  Multiple standards to virtualize of several of these components.
  • 18.
    Virtualization of NICs For multiple Vms on the system, each VM needs its own Virtual NIC.  One way to solve above problem is by using hypervisor software, that provides virtual CPU and also implements as many virtual NICs as VMs.  vNICs  vSWITCH  pNIC  pSwitch  VNIC implementation has different standard approaches.
  • 19.
    Virtualization of switches Ethernet switch has 32 – 128 ports  Number of physical machines that needs to be connected on L2 network is larger  Several layers of switches need to be used to form L2 network
  • 20.
    Virtual LANs inCLOUDS  Problem in cloud: Multiple VMs in a single physical machine may belong to different clients and need to be in different VLAN  Each VLAN span several data centers  Solution : VXLAN, NVGRE, STT
  • 21.
    Software Defined Networking SDN is latest revolution in networking innovations.  SDN consists of four innovations 1) Separation of control plane and data plane 2) Centralization of control plane 3) Programmability of control plane 4) Standardization of application programming interface(APIs)
  • 22.
    Separation of controland data planes  Networking protocols are arranged in different planes : data, control and management  Data plane : consists of all messages that are generated by users.  Control plane : Deals with transport of all the messages in data plane. Generates routing tables by using different routing protocols .  Management Plane: Keeps track of traffic statistics and states of various network equipment.  Key innovation of SDN is separation of Control and data plane.  Control logic is separated and implemented in controller that prepares forwarding table  This reduces the complexity and cost of the switches significantly.
  • 23.
    Centralization of controlplane  Centralization was considered bad thing until few years ago  Now it considered as good for good reason  Centralization of control makes sensing the state and adjusting the control dynamically based on state changes much faster than distributed protocols  Standby controllers can be used to take over in case of failures of the main controller
  • 24.
    Programmable control plane It is easy for the network manager to implement control changes by simply changing the control program.  The programmable control plane is the most important aspect of SDN  Programmable control plane allows the network to be divided in to several virtual networks with different policies and yet resides on shared hardware.
  • 25.
    Standardized APIs  SDNconsists of centralized control plane with  Southbound API for communication with hardware infrastructure  Northbound API is for communication with network applications
  • 26.
    SDN impact andfuture  SDN is expected to make network programmable and easily partitionable and virtualizable  These features are required for cloud computing where network infrastructure is shared with number of competing entities  SDN is expected to reduce both capital expenditure and operational expenditure  Network of tomorrow is more programmable than today
  • 27.
    Open Application deliveryusing SDN  Current SDN based efforts are restricted to L3 and below network traffic  It may be expanded to L3 and above layer network traffic management  Application traffic management involves application deployment and delivery policies.  Application service is replicated over multiple hosts and may be partitioned for improved performance.
  • 28.
    Problem statement  Mostapplications need to serve global audience  Needs servers all over the world  Cloud services provides multiple computing and storage facilities  Problem is routing using ASP’s policies in a very dynamic multi cloud environment is not possible  Since ISP’s offer no service to dynamically route messages
  • 29.
  • 30.
  • 31.
    Network Virtualization andresource description in software defined wireless Networks By Qianru zhou, Cheng-Xiang wang, Stephen Mclaughlin, and xiaotian zhou
  • 32.
    Network virtualization andresource description in software defined wireless networks  Challenges in wireless networks  Overview of existing SDWN architectures  Network description based on RDF  SDWN Architecture with resource description and ontologies
  • 33.
    Challenges in wirelessnetworks  SDN virtualizes the network architecture and isolates data control traffic.  The design of wireless network architecture challenging  Must deal with physical restrictions caused by fast changing nature of wireless channels  Server virtualization of wireless networks is also more challenging as it has to satisfy the requirements of both coherence and hardware isolation
  • 34.
    SDWN Virtualization Architecture SDWN is about making decisions on how a connection or flow is transmitted across network  SDWN is to split data and control plane  Most widely used protocol is open flow  It configure network elements  Provides open protocol to program the flow table in different switches and routers
  • 35.
    Architecture designs ofSDWN  Current SDWN research focuses on network architecture  Existing designs often focus on different positions  Route flow focuses on IP routing services  FlowVisor and FlowN concentrates on slicing the network physical infrastructure .  OpenRoads was proposed with the intention to replace present WIFI networks
  • 36.
  • 37.
    Information model  InSDWN, information model is the fundamental element  Information model describes all resources of network  This information model is foundation of network virtualization  It describes both physical layer infrastructure and visualization.  Information model should be Technology independent, reusable, easily extensible and linkable to other existing model.
  • 38.
    Semantic technology  Semanticweb all the information and services can be understood and used both by humans and computers.  Semantic web is composed of three elements  Metadata, RDF, and ontology  Metadata is the data about data  RDF is a standard about making statements about resources  Ontology is also known as vocabulary, describes set of classes and relationship between classes  In network description ontology describes a set of nodes and relation between them.
  • 39.
    Network semantic ontologyapplications  Until now network semantic ontology languages proposed are numberless  These languages have different grammar, different parameters and different specificities of application  A universally accepted language that describes the resources of SDN is not has ben proposed  Due to complicated and variable wireless channels environment and emerging new technologies building ontology for wireless networks is arduous task
  • 40.
    Performance evaluation ofontology  The ontology evaluation is the process to determine which resources the ontology defines correctly/incorrectly and those it does not define.  The criteria for performance evaluation are : 1) Consistency 2) Completeness 3) Conciseness 4)Expandability 5) Sensitiveness
  • 41.
    SDWN architecture withresource description function module
  • 42.
    Resource ontology  Withthis ontology we can express resources in RDF triples  “BaseStation A” “has Antenna” “Antenna 1”
  • 43.
    Qos ontology  Wecan describe Qos resources in RDF  “Node A” “hasMSTR” “10Mb/s”
  • 44.
  • 45.
  • 46.
    References  Michael Armbrust,Armando Fox, Rean Griffith, Anthony D. Joseph, Randy Katz, Andy Konwinski, Gunho Lee, David Patterson, Ariel Rabkin, Ion Stoica, and Matei Zaharia. 2010. A view of cloud computing. Commun. ACM 53, 4 (April 2010), 50-58. DOI=http://dx.doi.org/10.1145/1721654.1721672  R. Jain and S. Paul, "Network virtualization and software defined networking for cloud computing: a survey," in IEEE Communications Magazine, vol. 51, no. 11, pp. 24-31, November 2013.doi: 10.1109/MCOM.2013.6658648  Q. Zhou, C. X. Wang, S. McLaughlin and X. Zhou, "Network virtualization and resource description in software-defined wireless networks," in IEEE Communications Magazine, vol. 53, no. 11, pp. 110-117, November 2015. doi: 10.1109/MCOM.2015.7321979