Software Defined Networks Mooly Sagiv msagivacmorg 036407606 Tel Aviv University Thursday 1618 Physics 105 Monday 1416 Schrieber 317 Adviser Michael Shapira Hebrew University ID: 586529
Download Presentation The PPT/PDF document "Reasoning about" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Reasoning about Software Defined Networks
Mooly Sagivmsagiv@acm.org03-640-7606Tel Aviv UniversityThursday 16-18 (Physics 105)Monday 14-16 Schrieber 317Adviser: Michael ShapiraHebrew University
http://www.cs.tau.ac.il/~msagiv/courses/rsdn.htmlSlide2
ContentChallenges in SDNs
Programming Language AbstractionsProgramming Language TechniquesProgram Language Tools Other useful tools Slide3
Challenges in SDNProgramming complexity
ReliabiltySlide4
The Internet: A Remarkable Story
Tremendous successFrom research experiment to global infrastructureBrilliance of under-specifyingNetwork: best-effort packet deliveryHosts: arbitrary applicationsEnables innovation in applicationsWeb, P2P, VoIP, social networks, virtual worldsBut, change is easy only at the edge… 4Slide5
Inside the ‘Net: A Different Story…
Closed equipmentSoftware bundled with hardwareVendor-specific interfacesOver specifiedSlow protocol standardizationFew people can innovateEquipment vendors write the codeLong delays to introduce new features5
Impacts performance, security, reliability, cost…Slide6
Do We Need Innovation Inside?
6Many boxes (routers, switches, firewalls, …), with different interfaces.Slide7
How Hard are Networks to Manage?
Operating a network is expensiveMore than half the cost of a networkYet, operator error causes most outagesBuggy software in the equipmentRouters with 20+ million lines of codeCascading failures, vulnerabilities, etc.The network is “in the way”Especially a problem in data centers… and home networks7Slide8
Creating Foundation for Networking
A domain, not a disciplineAlphabet soup of protocolsHeader formats, bit twiddlingPreoccupation with artifactsFrom practice, to principlesIntellectual foundation for networkingIdentify the key abstractions… and support them efficientlyTo build networks worthy of society’s trust8Slide9
Rethinking the “Division of Labor”
9Slide10
Traditional Computer Networks
10
Data plane:
Packet streaming
Forward, filter, buffer, mark,
rate-limit, and measure packetsSlide11
Traditional Computer Networks
11
Track topology changes, compute routes, install forwarding rules
Control plane:
Distributed algorithmsSlide12
Traditional Computer Networks
12
Collect measurements and configure the equipment
Management plane:
Human time scaleSlide13
Shortest-Path RoutingManagement
: set the link weightsControl: compute shortest pathsData: forward packets to next hop13
1
1
3
1
1Slide14
Shortest-Path RoutingManagement
: set the link weightsControl: compute shortest pathsData: forward packets to next hop14
1
1
3
1
1Slide15
Inverting the Control PlaneTraffic engineering
Change link weights… to induce the paths… that alleviate congestion15
5
1
3
1
1Slide16
Avoiding Transient AnomaliesDistributed protocol
Temporary disagreement among the nodes… leaves packets stuck in loopsEven though the change was planned!16
1
5
1
3
1
1Slide17
Death to the Control Plane!Simpler management
No need to “invert” control-plane operationsFaster pace of innovationLess dependence on vendors and standardsEasier interoperabilityCompatibility only in “wire” protocolsSimpler, cheaper equipmentMinimal software17Slide18
Software Defined Networking (SDN)
18
API to the data plane
(e.g., OpenFlow)
Logically-centralized control
Switches
Smart,
slow
Dumb,
fastSlide19
OpenFlow Networks
19Slide20
Data-Plane: Simple Packet HandlingSimple packet-handling rules
Pattern: match packet header bitsActions: drop, forward, modify, send to controller Priority: disambiguate overlapping patternsCounters: #bytes and #packets20src=1.2.*.*, dest=3.4.5.* drop src = *.*.*.*, dest=3.4.* forward(2)3. src=10.1.2.3, dest=*.*.*.* send to controllerSlide21
Controller: Programmability
21Network OSApp #1App #2
App #3
Events from switches
Topology changes,
Traffic statistics,
Arriving packets
Commands to switches
(Un)install rules,
Query statistics,
Send packetsSlide22
OpenFlow in the Wild
Open Networking FoundationCreating Software Defined Networking standardsGoogle, Facebook, Microsoft, Yahoo, Verizon, Deutsche Telekom, and many other companiesCommercial OpenFlow switchesHP, NEC, Quanta, Dell, IBM, Juniper, …Network operating systemsNOX, Beacon, Floodlight, Nettle, ONIX, POX, FreneticNetwork deploymentsEight campuses, and two research backbone networksCommercial deployments22Slide23
Dynamic Access Control
Inspect first packet of each connectionConsult the access control policyInstall rules to block or route traffic23Slide24
Seamless Mobility/Migration
See host sending traffic at new locationModify rules to reroute the traffic24Slide25
Example Applications
Dynamic access controlSeamless mobility/migrationServer load balancingUsing multiple wireless access pointsEnergy-efficient networkingAdaptive traffic monitoringDenial-of-Service attack detectionNetwork virtualization25See http://www.openflow.org/videos/Slide26
Challenges of Programming Software Defined Networks
26Slide27
Programming OpenFlow Networks
OpenFlow makes programming possibleNetwork-wide view at controllerDirect control over data planeThe APIs do not make it easyLow level of abstractionChallengesCompositionConcurrencyCorrectnessTesting27Controller
SwitchesSlide28
A Simple example: FirewallA switch connected to two kind of hosts
Trusted hosts via port 1Untrusted hosts via port 2Trusted hosts can freely send packets to untrusted hostsAn unstrusted host can only send to a trusted destination which previously sent messages to this hostSlide29
Firewall
1
2Slide30
Firewall Controller Pseudo-code
rel trusted(SW, HO) packetIn(s, p, 1) # packets from trusted hosts send(s, p, 2) # forward the packet to untrusted hosts trusted.insert(s, p.dst) # insert the target of p into trusted controller memory ft.insert(s, p, 1, 2) # insert a per-flow rule to forward future packets packetIn(s, p, 2) -> # packets from untrusted hosts if trusted(s, p.src) then { send(s, p, 1) # forward the packet to trusted hosts ft.insert
(s, p, 2, 1) # insert a per-flow rule to forward future packets
}Slide31
Firewall Controller Pseudo-code(2)
packetIn(s, p, 1) # packets from trusted hosts send(s, p, 2) # forward the packet to untrusted hosts ft.insert(s, src:p.src, 1, 2) # insert a general rule to forward future packets ft.insert(s, dst:p.dst, 2, 1) # allow future packets from 2Slide32
A Learning Switch
Ttwo hosts (A & B) An OpenFlow switch with 3 portsHost A is connected to port 1and Host B is connected to port 2Gradually install forwarding rulesUpdate upon relocationHost ASwitch1
2
Host BSlide33
‘A’ sends a message to ‘B’
Host ASwitch12
Host B
TCP
syn
dst
=BSlide34
Forward to the Controller
Host ASwitch12
Host B
TCP
syn
dst
=B
3
send TCP
syn
dst
=B on port 2
send TCP
syn
dst
=B on port 3
learn that A is connected via port 1Slide35
‘B’ sends a message to ‘A’
Host ASwitch12
Host B
TCP
ack
dst
=ASlide36
Forward to the Controller
Host ASwitch12
Host B
TCP
ack
dst
=A
3
send TCP
ack
dst
=A on port 1
learn that B is connected via port 2
Install a rule to forward packets from B to A on port 1Slide37
‘A’ sends another message to ‘B’
Host ASwitch12
Host B
dst
=BSlide38
Forward to the Controller
Host ASwitch12
Host B
dst
=B
3
Send
dst
=B on port 2
Install a rule to forward packets from A to B to port 2Slide39
Learning Switch Pseudo-code
rel connected (SW, PR, HO)PacketIn(s, p, e) -> connect.insert (s, e, p.src) if connect(s, o, p.dst) then { send (s, p, o) ft.insert(s, p, e, o) } else foreach o in {1, 2, 3} – p # Flood send (s, p, o) Slide40
Reasoning about Programs
DebuggingTestingModel checkingProgramming language supportAbstractionCompositionEase of useProgram verificationAbstractionSlide41
Seminar BenefitsA cool topic
ReasoningCritically read an articleLearn to present an articleSlide42
Seminar RequirementsCompilers
Read an article (2 weeks)Prepare presentation (1 week)Participate in lectures Slide43
Tentative Schedule
October 24Michael Shapira Introduction to SDNOctober 31Mooly SagivIntroduction to Program ReasoningNovember 7??November 14??