Chapter 2 discusses the evolution of Software Defined Networks (SDN), highlighting changes in switches and control planes to meet increasing network demands. It emphasizes the shift towards centralized control, simplifying network management and reducing operational costs, while fostering innovation through open software environments. The chapter also addresses current inadequacies in networking and the need for scalable, automated solutions in data centers.

1. Chapter 2 1
Software Defined Networks
Chapter 2
Why SDN?

2. Chapter 2 Page 2
These slides have been prepared to accompany the book “Software Defined Networks – A Comprehensive Approach” Dr. Paul
Goransson, Chuck Black, and Timothy Culver. The authors and publisher Morgan Kaufman allow lecturers and public and private
universities the right to modify these slides for their own use when the book is being used for a class. Since the SDN marketplace
is changing rapidly, the accompanying PowerPoints and related classroom material will be updated every few months

3. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Chapter 2
• Early days – Simple Forwarding and Routing via
Software
• Network device developers and standards focused on
autonomous and independence
– Networks small w/ large shared domains
– Plug and play devices
– Manual configuration of devices
Page 3

4. Chapter 2
Why SDN?
Page 4
Evolution of Switches and Control Planes
Developers  implemented distributed environment with intelligence in each
device
• Coordination between devices  Collective decisions
• Goals
• Simplicity
• Ease of use
• Automatic Recovery

5. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Distributed Intelligence in L2 and L3
• Spanning Tree Protocol (STP) – IEEE 802.1D
– a.k.a. Transparent Bridging
– Enforces a hierarchy on the network
» Convergence latency
• Rapid Spanning Tree Protocol (RSTP) – IEEE 802.1D-2004
– Improves latency, but not deployed
Page 5

6. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Distributed Intelligence in L2 and L3
• Shortest Path Bridging (SPB)
– Multipath – via collaborative and distributed calculation of
shortest and most efficient paths
– Layer 2
Page 6

7. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Distributed Intelligence in L2 and L3
• RIP, BGP, OSPF, and IS-IS
– Chapter 2
» Layer 3 requires the cooperation between devices
• Knowledge of which routers are attaching to which
subnets in a network.
Page 7

8. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Distributed Intelligence in L2 and L3
• Routing functions move into hardware (see below)
– Applications Specific Integrated Circuits (ASICs)
– Fully Programmable Gate Arrays (FPGA’s)
– Ternary Content Addressable Memories (TCAMs)
Page 8
Forwarding decisions at line rate
1 GB 10 GB 40 GB
Speed

9. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Hardware Forwarding and Control in Software
• The network device evolution we have recounted thus
far has yielded the following current situation:
– Bridging (layer two forwarding).
– Routing (layer three forwarding).
– Advanced filtering and prioritization.
– Control
Page 9

10. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Growing Need for Simplicity
• Network Devices
» Why?
• Need for autonomous and independent design
• Increased intelligence
• Complicated handshakes
• Network Management
– Primitive tools like SNMP and CLI
Page 10
Simple Complex
Time

11. Chapter 2
Why SDN?
• Evolution of Switches and Control Planes
– Moving Control Off the Device
• Core of SDN  Move control software off the device and into a
centrally located compute resource
– Why?
» Central device can see entire network and make optimal decisions
– SDN attempts to segregate network activities in the
following manner:
• Forwarding, filtering, and prioritization.
• Control
• Application
Page 11
Here

12. Chapter 2 12

13. Chapter 2
Why SDN?
• Cost
– Increased Cost of Development
• Increasing demands on devices
• Factors
– Open Source Software
– COTS (common-off-the-shelf ASICs)
Page 13
Complicated
Control
Plane
Reuse of
common
frameworks
Rival and
surpass
proprietary
Requires Re-engineering

14. Chapter 2
Why SDN?
• Cost
– Closed Environments Encourage Vendor Lock-in
• Standards and Competitive Advantage
– Complexity and Resistance to Change
• “If its not broke, don’t fix it”
– Increased Cost of Operating the Network
• OPEX growing relative to CAPEX
– Why? Larger and more complex networks!
– SDN will lower OPEX
Page 14

15. Chapter 2
Why SDN?
• SDN Implications for Research and Innovation
– Status Quo Benefits Incumbent Vendors
• Few competitors
• Barriers to market entry due to costs
– Parallel to SaaS vendors (e.g. Force.com)
– SDN Promotes Research & Innovation
• Open software environments promote rapid pace of
advancement
– Examples: Linux, KVM, Xen, MySQL, Postgres
• Hardware commoditization and Openness  Innovation
Page 15

16. Chapter 2
Why SDN?
• Data Center Innovation
– Compute and Storage Virtualization
• It’s old! Circa 1972 @ IBM
– Network domain has not been innovating at the
same pace.
Page 16

17. Chapter 2
Physical Coordination
of Links and NICs
Physical Coordination
of Links and NICs
Why SDN?
• Inadequacies in Networks Today
– Growing need for network to respond to frequent
and immediate changes
– VM changes take minutes!
Page 17
Work Orders
Coordination
between server and
Network Admins
Logical Coordination
of Links and NICs
Physical Coordination
of Links and NICs
Physical Coordination
of ToR Switches
Network
Changes
Take
DAYS

18. Chapter 2
Why SDN?
Page 18

19. Chapter 2
Why SDN?
• Data Center Needs
Scalability
Network Virtualization
Automation
Multi-tenancy
Multi-pathing
Page 19
Limits of MAC and
number of VLANs
problematic
Dynamically
instantiate networks
as well as tear down
Abstract over
physical network
“anytime anywhere”
Each tenant has their
own virtual network
that they manage
Need to have
multiple paths

20. Chapter 2
Why SDN?
Page 20