3.2. OS model: We all know about OS model as shown in figure (3). It has three basic layers:
• Applications
• OS
• Hardware
The OS can manage hardware for many application. OS separate between hardware and the applications giving advantage to programmers to innovate.
3. Control plane and Data plane separation
3.3. SDN separation of planes: SDN architecture looks like OS model as shown in figure (4). It consists of three layers:
• Application layer: we can run here many kinds of application such that:
Cloud orchestration:
e.g. OpenStack, CloudStack…etc.
Business applications
SDN applications
…etc.
• Control layer: it is one logically centralized controller.
• Forwarding layer: it is many forwarding devices. Here we call tradition …show more content…
SDN controller communicates with forwarding devices using open interfaces (e.g. OpenFlow).
3.4. SDN misconceptions:
• The first misconception about the controller is that it must be physically centralized. In reality it is not required for example Onix and ONOS (OpenFlow OS) has physically distributed controllers (but logically centralized).
3. Control plane and Data plane separation
• The second misconception is that the first packet must go to controller for handling. Truly some past systems like Ethane worked like that for some security policies but it is not required in all systems.
• The third misconception is that SDN is OpenFlow. In fact, OpenFlow is a popular protocol but it is not the only one.
3.5. Why separation?
• Evolution of applications (software) is independent of hardware.
• More flexibility: can introduce new services more easily. No need to configure individual hardware.
• Increase innovation: not dependent on vendors and long standardization.
• Simpler management: control from high level control program.
• Fast adapt to changes in network world.
• Simpler and cheaper hardware.
• Increase reliability and security.
• Better end-user …show more content…
Commercial SDN controllers: Commercial SDN controllers also exist such that:
• Cisco onePK controller
• HP Virtual Application Networks (VAN) SDN Controller…etc.
But we want to take a look on a heavy industry that standardize a new model which will make revolution in the market of SDN controllers.
OpenDaylight: More than 52 corporations we listed some below: They intended to build one single platform based on open source for SDN/NVF (Network Functions Virtualization) platform. It will not only support OpenFlow but they have more innovation and intelligent work to do. References
1. Course: Software Defined Networking by Nick Feamster, Princeton University. https://www.coursera.org/course/sdn1 http://noise.gatech.edu/classes/cs8803sdn/fall2014/
2. Books:
• “SDN: Software Defined Networks” by Thomas D. Nadeau and Ken Gray, O’Reilly 2013
• “Network Innovation through OpenFlow and SDN Principles and Design” Edited by FEI HU, CRC Press 2014
3. Papers:
• “The Road to SDN: An Intellectual History of Programmable Networks” by Nick Feamster, Jennifer Rexford, and Ellen Zegura
• “OpenFlow: Enabling Innovation in Campus Networks” by Nick McKeown Stanford University March 14, 2008
4.
• Applications
• OS
• Hardware
The OS can manage hardware for many application. OS separate between hardware and the applications giving advantage to programmers to innovate.
3. Control plane and Data plane separation
3.3. SDN separation of planes: SDN architecture looks like OS model as shown in figure (4). It consists of three layers:
• Application layer: we can run here many kinds of application such that:
Cloud orchestration:
e.g. OpenStack, CloudStack…etc.
Business applications
SDN applications
…etc.
• Control layer: it is one logically centralized controller.
• Forwarding layer: it is many forwarding devices. Here we call tradition …show more content…
SDN controller communicates with forwarding devices using open interfaces (e.g. OpenFlow).
3.4. SDN misconceptions:
• The first misconception about the controller is that it must be physically centralized. In reality it is not required for example Onix and ONOS (OpenFlow OS) has physically distributed controllers (but logically centralized).
3. Control plane and Data plane separation
• The second misconception is that the first packet must go to controller for handling. Truly some past systems like Ethane worked like that for some security policies but it is not required in all systems.
• The third misconception is that SDN is OpenFlow. In fact, OpenFlow is a popular protocol but it is not the only one.
3.5. Why separation?
• Evolution of applications (software) is independent of hardware.
• More flexibility: can introduce new services more easily. No need to configure individual hardware.
• Increase innovation: not dependent on vendors and long standardization.
• Simpler management: control from high level control program.
• Fast adapt to changes in network world.
• Simpler and cheaper hardware.
• Increase reliability and security.
• Better end-user …show more content…
Commercial SDN controllers: Commercial SDN controllers also exist such that:
• Cisco onePK controller
• HP Virtual Application Networks (VAN) SDN Controller…etc.
But we want to take a look on a heavy industry that standardize a new model which will make revolution in the market of SDN controllers.
OpenDaylight: More than 52 corporations we listed some below: They intended to build one single platform based on open source for SDN/NVF (Network Functions Virtualization) platform. It will not only support OpenFlow but they have more innovation and intelligent work to do. References
1. Course: Software Defined Networking by Nick Feamster, Princeton University. https://www.coursera.org/course/sdn1 http://noise.gatech.edu/classes/cs8803sdn/fall2014/
2. Books:
• “SDN: Software Defined Networks” by Thomas D. Nadeau and Ken Gray, O’Reilly 2013
• “Network Innovation through OpenFlow and SDN Principles and Design” Edited by FEI HU, CRC Press 2014
3. Papers:
• “The Road to SDN: An Intellectual History of Programmable Networks” by Nick Feamster, Jennifer Rexford, and Ellen Zegura
• “OpenFlow: Enabling Innovation in Campus Networks” by Nick McKeown Stanford University March 14, 2008
4.