Praveen Tammana† Rachit Agarwal‡ Myungjin Lee†

Slides:



Advertisements
Similar presentations
Interconnection Networks: Flow Control and Microarchitecture.
Advertisements

Towards Software Defined Cellular Networks
Compiling Path Queries in Software-Defined Networks Srinivas Narayana Jennifer Rexford and David Walker Princeton University.
Practical and Incremental Convergence between SDN and Middleboxes 1 Zafar Qazi, Cheng-Chun Tu, Luis Chiang Vyas Sekar Rui Miao Minlan Yu.
Models and techniques for verification of Software Defined Networks
VCRIB: Virtual Cloud Rule Information Base Masoud Moshref, Minlan Yu, Abhishek Sharma, Ramesh Govindan HotCloud 2012.
Nanxi Kang Princeton University
Joining LANs - Bridges. Connecting LANs 4 Repeater –Operates at the Physical layer no decision making, processing signal boosting only 4 Bridges –operates.
Fundamentals of Computer Networks ECE 478/578 Lecture #13: Packet Switching (2) Instructor: Loukas Lazos Dept of Electrical and Computer Engineering University.
Incremental Consistent Updates Naga Praveen Katta Jennifer Rexford, David Walker Princeton University.
Consensus Routing: The Internet as a Distributed System John P. John, Ethan Katz-Bassett, Arvind Krishnamurthy, and Thomas Anderson Presented.
USING PACKET HISTORIES TO TROUBLESHOOT NETWORKS Presented by: Yi Gao Emnets Seminar
Towards Virtual Routers as a Service 6th GI/ITG KuVS Workshop on “Future Internet” November 22, 2010 Hannover Zdravko Bozakov.
Scalable Flow-Based Networking with DIFANE 1 Minlan Yu Princeton University Joint work with Mike Freedman, Jennifer Rexford and Jia Wang.
Flowspace revisited OpenFlow Basics Flow Table Entries Switch Port MAC src MAC dst Eth type VLAN ID IP Src IP Dst IP Prot L4 sport L4 dport Rule Action.
1 Link Layer & Network Layer Some slides are from lectures by Nick Mckeown, Ion Stoica, Frans Kaashoek, Hari Balakrishnan, and Sam Madden Prof. Dina Katabi.
Interference-aware QoS Routing (IQRouting) for Ad-Hoc Networks Rajarshi Gupta, Zhanfeng Jia, Teresa Tung, and Jean Walrand Dept of EECS, UC Berkeley Globecom.
A Scalable, Commodity Data Center Network Architecture.
Languages for Software-Defined Networks Nate Foster, Arjun Guha, Mark Reitblatt, and Alec Story, Cornell University Michael J. Freedman, Naga Praveen Katta,
Enabling Layer 2 Pathlet Tracing through Context Encoding in SDN Hui Zhang, Cristian Lumezanu, Junghwan Rhee, Nipun Arora, Qiang Xu, Guofei Jiang NEC Labs.
BUFFALO: Bloom Filter Forwarding Architecture for Large Organizations Minlan Yu Princeton University Joint work with Alex Fabrikant,
Not All Microseconds are Equal: Fine-Grained Per-Flow Measurements with Reference Latency Interpolation Myungjin Lee †, Nick Duffield‡, Ramana Rao Kompella†
Cellular Core Network Architecture
Formal checkings in networks James Hongyi Zeng with Peyman Kazemian, George Varghese, Nick McKeown.
Fast Failover for Control Traffic in Software-defined Networks Globecom 2012 Neda B. & Ying Z. Presented by: Szu-Ping Wang.
VeriFlow: Verifying Network-Wide Invariants in Real Time
A Scalable, Commodity Data Center Network Architecture Jingyang Zhu.
DARD: Distributed Adaptive Routing for Datacenter Networks Xin Wu, Xiaowei Yang.
L11: Link and Network layer Dina Katabi Spring Some slides are from lectures by Nick Mckeown, Ion Stoica, Frans Kaashoek,
1 Countering DoS Through Filtering Omar Bashir Communications Enabling Technologies
CellSDN: Software-Defined Cellular Core networks Xin Jin Princeton University Joint work with Li Erran Li, Laurent Vanbever, and Jennifer Rexford.
Coping with Link Failures in Centralized Control Plane Architecture Maulik Desai, Thyagarajan Nandagopal.
Header Space Analysis: Static Checking for Networks Broadband Network Technology Integrated M.S. and Ph.D. Eun-Do Kim Network Standards Research Section.
BUFFALO: Bloom Filter Forwarding Architecture for Large Organizations Minlan Yu Princeton University Joint work with Alex Fabrikant,
Flow OAM Requirements Janardhanan Pathangi Balaji Venkat Venkataswami DELL Richard Groves – Microsoft Peter Hoose – Facebook
COMP8330/7330/7336 Advanced Parallel and Distributed Computing Communication Costs in Parallel Machines Dr. Xiao Qin Auburn University
Network Virtualization Ben Pfaff Nicira Networks, Inc.
Ready-to-Deploy Service Function Chaining for Mobile Networks
SDN and Security Security as a service in the cloud
Xin Li, Chen Qian University of Kentucky
SDN challenges Deployment challenges
CIS 700-5: The Design and Implementation of Cloud Networks
SDN Network Updates Minimum updates within a single switch
University of Maryland College Park
The DPIaaS Controller Prototype
Heitor Moraes, Marcos Vieira, Italo Cunha, Dorgival Guedes
Martin Casado, Nate Foster, and Arjun Guha CACM, October 2014
Hydra: Leveraging Functional Slicing for Efficient Distributed SDN Controllers Yiyang Chang, Ashkan Rezaei, Balajee Vamanan, Jahangir Hasan, Sanjay Rao.
Efficient Round-Trip Time Monitoring in OpenFlow Networks
Revisiting Ethernet: Plug-and-play made scalable and efficient
NOX: Towards an Operating System for Networks
of Dynamic NFV-Policies
Intra-Domain Routing Jacob Strauss September 14, 2006.
Abstractions for Model Checking SDN Controllers
Dingming Wu+, Yiting Xia+*, Xiaoye Steven Sun+,
DDoS Attack Detection under SDN Context
Interconnection Network Design Lecture 14
NTHU CS5421 Cloud Computing
ECE 544 Project3 Team member: BIAO LI, BO QU, XIAO ZHANG 1 1.
Data and Computer Communications
IP Traceback Problem: How do we determine where malicious packet came from ? It’s a problem because attacker can spoof source IP address If we know where.
A Probabilistic Routing Protocol for Mobile Ad Hoc Networks
Local configuration global effect
2019/5/2 Using Path Label Routing in Wide Area Software-Defined Networks with OpenFlow ICNP = International Conference on Network Protocols Presenter:Hung-Yen.
Toward Self-Driving Networks
Toward Self-Driving Networks
Jennifer Rexford Princeton University
Elmo Muhammad Shahbaz Lalith Suresh, Jennifer Rexford, Nick Feamster,
Control-Data Plane Separation
Chapter 4: outline 4.1 Overview of Network layer data plane
Presentation transcript:

Praveen Tammana† Rachit Agarwal‡ Myungjin Lee† CherryPick: Tracing Packet Trajectory in Software-Defined Datacenter Networks Praveen Tammana† Rachit Agarwal‡ Myungjin Lee† †University of Edinburgh ‡UC Berkeley

Debugging datacenter networks SDN-enabled datacenters Flexible network management Control plane policies to data plane rules Debugging becomes more challenging Ensuring data plane conforms with control plane .

Packet trajectory tracing “Tracing” the path taken by the packet Scalability: Switch flow rules, packet header space Policy: All packets from 1 to 3 should go through 4 Checks whether packet followed route as defined in policy Helps to localize network problems Ex: misconfiguration, failures 4 5 6 mismatch Actual path : 1 - 6 - 3 1 2 3

Enabling packet trajectory tracing Challenge: Packets traversing non-shortest paths Idea 1: Append link IDs at each hop Insert ingress link ID Large packet header space as path length increases 5 User packet 4 6 Src Dst 1 2 3

Enabling packet trajectory tracing Challenge: Packets traversing non-shortest paths Idea 2: Maintain global path ID for each unique path Path id 1 Path id 2 # Paths in 48-ary fat-tree 4 hops : 576 6 hops: 1.3 million !! Need one flow rule for each unique path id # Switch flow rules ∝ # Paths that contain switch Src Dst

Handles wide range of queries Existing approaches Naïve approach Switches append ingress link ID into the packet header Large number of header bits PathletTracer [HotSDN’14] and PathQuery [HotSDN’14] Switches insert or update path information in packet header Large number of switch flow rules NetSight [NSDI’14] Logs packet details at each hop Need to collect large amount of data Topology agnostic Handles wide range of queries

CherryPick overview Exploits the structure in data center topologies (e.g., fat-tree) Observation: A small subset of links sufficient to represent any end-to-end path Scalability: Minimal number of flow rules irrespective to path length Packet header space requirement is close to state-of-the-art th

Selectively picking links 4 hops Pick ingress link id-1 Core Only one shortest path from Core to Dst Core Picked link Agg All 4 hops: Pick 1 link Agg User packet ToR Src Dst

Selectively picking links 6 hops id-1 Only one shortest path between selected links id-1 Picked link Pick ingress link id-2 All 6 hops: Pick 2 links Agg Picked link ToR Src Dst For more details, please refer to our paper

Evaluation results for 48-ary fat-tree ToR Switch Independent of path length Minimal number of switch flow rules Header space requirement is close to PathletTracer 107 105

Conclusion CherryPick Future work Simple and scalable packet trajectory tracing technique Demonstrate efficacy under fat-tree topology Feasibility: Trace all 4- and 6-hop paths using double tagging Low overhead: #OpenFlow rules = #Ports Future work Generalization to other datacenter topologies Tracing trajectories of packets that do not reach destination

Thank You

Backup-slides Spend time on main slide -> more animations, prepare Keep answer short Talk & animation should be on sync Screen wall far – don’t point General framework : Start making story and come to point directly

Selectively picking a minimum number of links id-1 Pick ingress link Pick ingress link Pick ingress link id-3 All 8 hops: Pick 3 links Src Dst id-2 Change diagram that embed id’s into pkt header (instead of showing just id’s) Add bullet points describing technique Title: Exploits topology Mention path reconstruction