SlideShare a Scribd company logo
Barista
An Event-centric NOS Composition Framework
for Software-Defined Networks
Jaehyun Nam1, Hyeonseong Jo1, Yeonkeun Kim1,
Phillip Porras2, and Vinod Yegneswaran2, and Seungwon Shin1
KAIST1 and SRI International2
Flexible Network Configurations
§ Software-defined networks (SDNs)
• Programmable control layer à increasing network agility
・Unified integration and management of diverse functions
• Deployment in various network environments
・E.g., enterprise, data center and cloud networks
§ Network operating systems (NOSs)
• The strategic control center of SDNs
・Management of network switches
and high-level applications
・Global network programmability
2
NOS Characteristics
§ A variety of NOSs
• Academia à Onix, Beacon, Rosemary, SE-Floodlight
• Open-source projects à
• Industry à
§ Specialized NOSs in different dimensions
• Specific purposes to design NOSs, resulting in architectural differences
• Examples
・Beacon for high performance
・ONOS and OpenDaylight for distributed scalability
・SE-Floodlight for enhanced security
3
Architectural Composition Challenge
§ What if we need a scalable and secure NOS?
• ONOS with scalability à high parallelism
• SE-Floodlight with security à centralized control
§ Architectural Gap between NOS design principles
• SE-Floodlight à ONOS
・Hard to inspect control flows due to direct comm. between components
• ONOS à SE-Floodlight
・Hard to achieve high performance due to centralized inspection mechanisms
・E.g., role-based authorization, flow rule conflict resolution
4
Barista
§ A novel architecture
• Enable flexible and customizable instantiations of NOSs
• Leverage both high parallelism and centralized control
§ Key Contributions
• Flexible component synthesis
・Accelerate rapid prototyping and integration of new NOS functionalities
• Dynamic event control
・Enable fine-grained control over events delivered to NOS components
◦ A diverse set of event types
◦ Dynamic chaining among components
◦ Policy-based event distribution
5
Barista System Overview
Operator
Operator’s
component
3-rd party
component
Barista
component
Barista (base framework)
Barista
Component pool
Component
management
Event
management
Command-line
interface
Admin
Network
Network
Security
Security
Network
Management
Network
Base
<Dynamic component event chaining>
Deploy
Pick-and-choose
Event
Worker
Event
Worker
Event
Worker
<Event handler>
Seq Seq
Parallel
Seq
Seq
Seq
Parallel
Event A
Event B
Event C
Centralized control
High
parallelism
6
Component
§ Implementation of a specific NOS function
§ Two types of components
⁃ General component
・Embedded as a part of control flows (e.g., OpenFlow engine)
⁃ Autonomous component
・Designed to take independent actions (e.g., Resource management)
§ Component pool
⁃ A set of components supporting the functionalities of contemporary NOSs
・Distributed scaling, performance improvement, security, failure isolation, etc.
7
Event
§ Message delivery mechanism among components
§ No direct communication between components
• All communications through the event handler
§ Enable high composability
8
Component #2
Component #5
Component #3
Component #n
Event handler
Component #1
Component #4
Event #1
Event #2
Event #3
Event #4
Component Portability
§ Execute the same code of a component either inside or outside
of the framework
§ Internal components
• Higher performance
• Limited flexibility
§ External components
• High flexibility and availability
• Lower performance
Barista framework
Source code
of a component
Wrapper library Source code
of a component
Event pipelines Control channel
9
Diverse Event Classes
§ Three types of event classes
• (i) Notification events, (ii) inter-component events, and (iii) meta events
Event queue Worker pool
C #1 C #2 C #3
Data
Push
Worker #1
Worker #2
Worker #n
Event chain
Component
Pop
(i) Notification events
(ii) Inter-component
events
Look up a receiver
Check event types
Return results
to the trigger
Triggered events Reply data
Push
Event broker
Update results
Statistics Base framework Component
(De)activate
Collect Meta events
(iii) Meta events
10
Dynamic Component Event Chaining
§ Two ways to deliver events to components
• Sequential vs. parallel delivery
§ Event sequence formulation
• Ordering components based on their role and permissions
• Evaluating which component can be processed in parallel or sequential
Control flow
integrity Flow mgmt OFP engine
Rule conflict
check
Security / r-x Security / r-x Mgmt / r-- Base / r--
Data
Role = {admin | security | network | management | base}
Perm = {r (read) | w (update data) | x (cut off its control flow)}
Role / Permission
Control flow
integrity Flow mgmt
OFP engine
Rule conflict
check
Should be in sequential Okay to be in parallel
11
Policy-based Event Distribution
§ Event handling policy for deployed components
§ Operator-defined policy (ODP)
• Datapath ID, in-port, protocol, source and dst IP addresses and ports
Event queues
pi
pj
pk
…
Pop
Add an ODP to the target component
Policy table Component
Handler
Check
DPID
Inport
Proto
Src IP
Dst IP
Sport
Dport
DPID1 DPID2
Any
IPv4
Src IP1 Src IP2
Dst IP range 1
Any
Any
Skip if not matched
12
Event Distribution Across Instances
§ Cluster component for event distribution
• Store triggered events into a distributed storage
• Keep polling new events in the distributed storage
§ Maintenance of logical sequence of events
• Eventual consistency support
§ Operator-defined event sharing
• Which instance listens to what kinds of events
• E.g., sharing ONLY switch and topology events with other instances
13
Implementation
§ Base framework and a broad set of components
• 17K lines of C code and Python script
• Maria DB, Galera Cluster for event distribution across instances
§ Open-source project
• All implementation details will be provided at
https://github.com/sdx4u/barista
14
Distributed and Secure NOS
§ Need a distributed and secure NOS?
• No controller that currently supports scalability and security together
§ NOS composition with the Barista component pool
• Pick and choose the following components from the Barista pool
⁃ “cluster, role-based authorization, component access control, and flow rule conflict resolution”
0
200K
400K
600K
800K
1.0M
1.2M
1 2 4 8 16 32 64 128
Throughput
(responses/s)
The number of switches
Barista ONOS SE-Floodlight
15
Policy-based Event Distribution
§ Selective event distribution based on ODPs
⁃ Forwarding à DPID:!2 / DPID:2; Port:1,2; Proto:LLDP
⁃ VNF manger à DPID:2
SDN controller
<Current Approach> <New Approach>
Barista NOS
Forwarding
VNF manager Forwarding
VNF manager
SW#1 SW#3
SW#2
NF
NF NF
Unified network view Per-component network view
SW#2
NF
NF NF
SW#1 SW#3
SW#2
Selectively deliver events
Deliver all events
VNF events Non-VNF events
2
1 2
1
16
Dynamic Component Activation
§ Monitor event statistics in the event handler
⁃ Event volume, component level statistics regarding event consumption and production
⁃ Event distribution statistics
§ Dynamically trigger meta events defined by operators
based on event statistics
SDN controller
<Current Approach> <Barista Approach>
Barista NOS
DDoS preventer
(1-2) Run it all the time
Operator
DDoS attack
happens!
Event handler DDoS preventer
(0) Define a meta event
“Activate the DDoS preventer
if # of new flows > 1,000 pec second”
(2) Activate the DDoS
preventer by itself
(1) # of new flows > 1,000 per sec
(1-1) Activate the DDoS preventer in manual
17
Conclusion
§ Contemporary NOSs à specialized NOSs
• Optimized controller design across one or a few dimensions
• Fundamental differences in architectural trade-offs needed
to support competing demands
§ Barista, A framework to build customized NOSs
suited for a variety of requirements from network operators
• Support flexible composition of functionalities in contemporary NOSs
• Dynamic control of associated events for deployed functionalities
18
Github: https://github.com/sdx4u/barista
Thank you for listening

More Related Content

Similar to Barista: Event-centric NOS Composition Framework for SDN

Lesson_08_Continuous_Monitoring.pdf
Lesson_08_Continuous_Monitoring.pdfLesson_08_Continuous_Monitoring.pdf
Lesson_08_Continuous_Monitoring.pdf
Minh Quân Đoàn
 
SDN Security Talk - (ISC)2_3
SDN Security Talk - (ISC)2_3SDN Security Talk - (ISC)2_3
SDN Security Talk - (ISC)2_3Wen-Pai Lu
 
SplunkApplicationLoggingBestPractices_Template_2.3.pdf
SplunkApplicationLoggingBestPractices_Template_2.3.pdfSplunkApplicationLoggingBestPractices_Template_2.3.pdf
SplunkApplicationLoggingBestPractices_Template_2.3.pdf
TuynNguyn819213
 
Big Data Berlin v8.0 Stream Processing with Apache Apex
Big Data Berlin v8.0 Stream Processing with Apache Apex Big Data Berlin v8.0 Stream Processing with Apache Apex
Big Data Berlin v8.0 Stream Processing with Apache Apex
Apache Apex
 
Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...
Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...
Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...
Dataconomy Media
 
Event log analyzer by me
Event log analyzer by me Event log analyzer by me
Event log analyzer by me
ER Swapnil Raut
 
[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'
[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'
[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'
OpenStack Korea Community
 
Exploring the Final Frontier of Data Center Orchestration: Network Elements -...
Exploring the Final Frontier of Data Center Orchestration: Network Elements -...Exploring the Final Frontier of Data Center Orchestration: Network Elements -...
Exploring the Final Frontier of Data Center Orchestration: Network Elements -...
Puppet
 
Platform Observability and Infrastructure Closed Loops
Platform Observability and Infrastructure Closed LoopsPlatform Observability and Infrastructure Closed Loops
Platform Observability and Infrastructure Closed Loops
Liz Warner
 
Acceleration_and_Security_draft_v2
Acceleration_and_Security_draft_v2Acceleration_and_Security_draft_v2
Acceleration_and_Security_draft_v2Srinivasa Addepalli
 
RuSIEM overview (english version)
RuSIEM overview (english version)RuSIEM overview (english version)
RuSIEM overview (english version)
Olesya Shelestova
 
Why Serverless Flink Matters - Blazing Fast Stream Processing Made Scalable
Why Serverless Flink Matters - Blazing Fast Stream Processing Made ScalableWhy Serverless Flink Matters - Blazing Fast Stream Processing Made Scalable
Why Serverless Flink Matters - Blazing Fast Stream Processing Made Scalable
HostedbyConfluent
 
SDN Security: Two Sides of the Same Coin
SDN Security: Two Sides of the Same CoinSDN Security: Two Sides of the Same Coin
SDN Security: Two Sides of the Same Coin
Zivaro Inc
 
Data center webinar_v2_1
Data center webinar_v2_1Data center webinar_v2_1
Data center webinar_v2_1Lancope, Inc.
 
Using Kubernetes to make cellular data plans cheaper for 50M users
Using Kubernetes to make cellular data plans cheaper for 50M usersUsing Kubernetes to make cellular data plans cheaper for 50M users
Using Kubernetes to make cellular data plans cheaper for 50M users
Mirantis
 
NetFlow Analyzer Training Part I: Getting the initial settings right
NetFlow Analyzer Training Part I: Getting the initial settings rightNetFlow Analyzer Training Part I: Getting the initial settings right
NetFlow Analyzer Training Part I: Getting the initial settings right
ManageEngine, Zoho Corporation
 
DEVNET-1164 Using OpenDaylight for Notification Driven Workflows
DEVNET-1164	Using OpenDaylight for Notification Driven WorkflowsDEVNET-1164	Using OpenDaylight for Notification Driven Workflows
DEVNET-1164 Using OpenDaylight for Notification Driven Workflows
Cisco DevNet
 
Software Innovations and Control Plane Evolution in the new SDN Transport Arc...
Software Innovations and Control Plane Evolution in the new SDN Transport Arc...Software Innovations and Control Plane Evolution in the new SDN Transport Arc...
Software Innovations and Control Plane Evolution in the new SDN Transport Arc...
Cisco Canada
 
Streaming real time data with Vibe Data Stream
Streaming real time data with Vibe Data StreamStreaming real time data with Vibe Data Stream
Streaming real time data with Vibe Data Stream
InformaticaMarketplace
 
Twelve Factor App
Twelve Factor AppTwelve Factor App
Twelve Factor App
Christ Ngantung
 

Similar to Barista: Event-centric NOS Composition Framework for SDN (20)

Lesson_08_Continuous_Monitoring.pdf
Lesson_08_Continuous_Monitoring.pdfLesson_08_Continuous_Monitoring.pdf
Lesson_08_Continuous_Monitoring.pdf
 
SDN Security Talk - (ISC)2_3
SDN Security Talk - (ISC)2_3SDN Security Talk - (ISC)2_3
SDN Security Talk - (ISC)2_3
 
SplunkApplicationLoggingBestPractices_Template_2.3.pdf
SplunkApplicationLoggingBestPractices_Template_2.3.pdfSplunkApplicationLoggingBestPractices_Template_2.3.pdf
SplunkApplicationLoggingBestPractices_Template_2.3.pdf
 
Big Data Berlin v8.0 Stream Processing with Apache Apex
Big Data Berlin v8.0 Stream Processing with Apache Apex Big Data Berlin v8.0 Stream Processing with Apache Apex
Big Data Berlin v8.0 Stream Processing with Apache Apex
 
Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...
Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...
Thomas Weise, Apache Apex PMC Member and Architect/Co-Founder, DataTorrent - ...
 
Event log analyzer by me
Event log analyzer by me Event log analyzer by me
Event log analyzer by me
 
[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'
[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'
[OpenStack Day in Korea 2015] Track 2-3 - 오픈스택 클라우드에 최적화된 네트워크 가상화 '누아지(Nuage)'
 
Exploring the Final Frontier of Data Center Orchestration: Network Elements -...
Exploring the Final Frontier of Data Center Orchestration: Network Elements -...Exploring the Final Frontier of Data Center Orchestration: Network Elements -...
Exploring the Final Frontier of Data Center Orchestration: Network Elements -...
 
Platform Observability and Infrastructure Closed Loops
Platform Observability and Infrastructure Closed LoopsPlatform Observability and Infrastructure Closed Loops
Platform Observability and Infrastructure Closed Loops
 
Acceleration_and_Security_draft_v2
Acceleration_and_Security_draft_v2Acceleration_and_Security_draft_v2
Acceleration_and_Security_draft_v2
 
RuSIEM overview (english version)
RuSIEM overview (english version)RuSIEM overview (english version)
RuSIEM overview (english version)
 
Why Serverless Flink Matters - Blazing Fast Stream Processing Made Scalable
Why Serverless Flink Matters - Blazing Fast Stream Processing Made ScalableWhy Serverless Flink Matters - Blazing Fast Stream Processing Made Scalable
Why Serverless Flink Matters - Blazing Fast Stream Processing Made Scalable
 
SDN Security: Two Sides of the Same Coin
SDN Security: Two Sides of the Same CoinSDN Security: Two Sides of the Same Coin
SDN Security: Two Sides of the Same Coin
 
Data center webinar_v2_1
Data center webinar_v2_1Data center webinar_v2_1
Data center webinar_v2_1
 
Using Kubernetes to make cellular data plans cheaper for 50M users
Using Kubernetes to make cellular data plans cheaper for 50M usersUsing Kubernetes to make cellular data plans cheaper for 50M users
Using Kubernetes to make cellular data plans cheaper for 50M users
 
NetFlow Analyzer Training Part I: Getting the initial settings right
NetFlow Analyzer Training Part I: Getting the initial settings rightNetFlow Analyzer Training Part I: Getting the initial settings right
NetFlow Analyzer Training Part I: Getting the initial settings right
 
DEVNET-1164 Using OpenDaylight for Notification Driven Workflows
DEVNET-1164	Using OpenDaylight for Notification Driven WorkflowsDEVNET-1164	Using OpenDaylight for Notification Driven Workflows
DEVNET-1164 Using OpenDaylight for Notification Driven Workflows
 
Software Innovations and Control Plane Evolution in the new SDN Transport Arc...
Software Innovations and Control Plane Evolution in the new SDN Transport Arc...Software Innovations and Control Plane Evolution in the new SDN Transport Arc...
Software Innovations and Control Plane Evolution in the new SDN Transport Arc...
 
Streaming real time data with Vibe Data Stream
Streaming real time data with Vibe Data StreamStreaming real time data with Vibe Data Stream
Streaming real time data with Vibe Data Stream
 
Twelve Factor App
Twelve Factor AppTwelve Factor App
Twelve Factor App
 

Recently uploaded

Designing for Privacy in Amazon Web Services
Designing for Privacy in Amazon Web ServicesDesigning for Privacy in Amazon Web Services
Designing for Privacy in Amazon Web Services
KrzysztofKkol1
 
Globus Compute Introduction - GlobusWorld 2024
Globus Compute Introduction - GlobusWorld 2024Globus Compute Introduction - GlobusWorld 2024
Globus Compute Introduction - GlobusWorld 2024
Globus
 
Using IESVE for Room Loads Analysis - Australia & New Zealand
Using IESVE for Room Loads Analysis - Australia & New ZealandUsing IESVE for Room Loads Analysis - Australia & New Zealand
Using IESVE for Room Loads Analysis - Australia & New Zealand
IES VE
 
Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...
Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...
Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...
Globus
 
Dominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdf
Dominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdfDominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdf
Dominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdf
AMB-Review
 
Large Language Models and the End of Programming
Large Language Models and the End of ProgrammingLarge Language Models and the End of Programming
Large Language Models and the End of Programming
Matt Welsh
 
Enhancing Research Orchestration Capabilities at ORNL.pdf
Enhancing Research Orchestration Capabilities at ORNL.pdfEnhancing Research Orchestration Capabilities at ORNL.pdf
Enhancing Research Orchestration Capabilities at ORNL.pdf
Globus
 
Understanding Globus Data Transfers with NetSage
Understanding Globus Data Transfers with NetSageUnderstanding Globus Data Transfers with NetSage
Understanding Globus Data Transfers with NetSage
Globus
 
Explore Modern SharePoint Templates for 2024
Explore Modern SharePoint Templates for 2024Explore Modern SharePoint Templates for 2024
Explore Modern SharePoint Templates for 2024
Sharepoint Designs
 
Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...
Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...
Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...
Shahin Sheidaei
 
Webinar: Salesforce Document Management 2.0 - Smarter, Faster, Better
Webinar: Salesforce Document Management 2.0 - Smarter, Faster, BetterWebinar: Salesforce Document Management 2.0 - Smarter, Faster, Better
Webinar: Salesforce Document Management 2.0 - Smarter, Faster, Better
XfilesPro
 
Visitor Management System in India- Vizman.app
Visitor Management System in India- Vizman.appVisitor Management System in India- Vizman.app
Visitor Management System in India- Vizman.app
NaapbooksPrivateLimi
 
How Recreation Management Software Can Streamline Your Operations.pptx
How Recreation Management Software Can Streamline Your Operations.pptxHow Recreation Management Software Can Streamline Your Operations.pptx
How Recreation Management Software Can Streamline Your Operations.pptx
wottaspaceseo
 
GlobusWorld 2024 Opening Keynote session
GlobusWorld 2024 Opening Keynote sessionGlobusWorld 2024 Opening Keynote session
GlobusWorld 2024 Opening Keynote session
Globus
 
Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...
Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...
Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...
Globus
 
SOCRadar Research Team: Latest Activities of IntelBroker
SOCRadar Research Team: Latest Activities of IntelBrokerSOCRadar Research Team: Latest Activities of IntelBroker
SOCRadar Research Team: Latest Activities of IntelBroker
SOCRadar
 
Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...
Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...
Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...
Hivelance Technology
 
TROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERROR
TROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERRORTROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERROR
TROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERROR
Tier1 app
 
How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?
How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?
How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?
XfilesPro
 
Developing Distributed High-performance Computing Capabilities of an Open Sci...
Developing Distributed High-performance Computing Capabilities of an Open Sci...Developing Distributed High-performance Computing Capabilities of an Open Sci...
Developing Distributed High-performance Computing Capabilities of an Open Sci...
Globus
 

Recently uploaded (20)

Designing for Privacy in Amazon Web Services
Designing for Privacy in Amazon Web ServicesDesigning for Privacy in Amazon Web Services
Designing for Privacy in Amazon Web Services
 
Globus Compute Introduction - GlobusWorld 2024
Globus Compute Introduction - GlobusWorld 2024Globus Compute Introduction - GlobusWorld 2024
Globus Compute Introduction - GlobusWorld 2024
 
Using IESVE for Room Loads Analysis - Australia & New Zealand
Using IESVE for Room Loads Analysis - Australia & New ZealandUsing IESVE for Room Loads Analysis - Australia & New Zealand
Using IESVE for Room Loads Analysis - Australia & New Zealand
 
Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...
Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...
Exploring Innovations in Data Repository Solutions - Insights from the U.S. G...
 
Dominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdf
Dominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdfDominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdf
Dominate Social Media with TubeTrivia AI’s Addictive Quiz Videos.pdf
 
Large Language Models and the End of Programming
Large Language Models and the End of ProgrammingLarge Language Models and the End of Programming
Large Language Models and the End of Programming
 
Enhancing Research Orchestration Capabilities at ORNL.pdf
Enhancing Research Orchestration Capabilities at ORNL.pdfEnhancing Research Orchestration Capabilities at ORNL.pdf
Enhancing Research Orchestration Capabilities at ORNL.pdf
 
Understanding Globus Data Transfers with NetSage
Understanding Globus Data Transfers with NetSageUnderstanding Globus Data Transfers with NetSage
Understanding Globus Data Transfers with NetSage
 
Explore Modern SharePoint Templates for 2024
Explore Modern SharePoint Templates for 2024Explore Modern SharePoint Templates for 2024
Explore Modern SharePoint Templates for 2024
 
Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...
Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...
Gamify Your Mind; The Secret Sauce to Delivering Success, Continuously Improv...
 
Webinar: Salesforce Document Management 2.0 - Smarter, Faster, Better
Webinar: Salesforce Document Management 2.0 - Smarter, Faster, BetterWebinar: Salesforce Document Management 2.0 - Smarter, Faster, Better
Webinar: Salesforce Document Management 2.0 - Smarter, Faster, Better
 
Visitor Management System in India- Vizman.app
Visitor Management System in India- Vizman.appVisitor Management System in India- Vizman.app
Visitor Management System in India- Vizman.app
 
How Recreation Management Software Can Streamline Your Operations.pptx
How Recreation Management Software Can Streamline Your Operations.pptxHow Recreation Management Software Can Streamline Your Operations.pptx
How Recreation Management Software Can Streamline Your Operations.pptx
 
GlobusWorld 2024 Opening Keynote session
GlobusWorld 2024 Opening Keynote sessionGlobusWorld 2024 Opening Keynote session
GlobusWorld 2024 Opening Keynote session
 
Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...
Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...
Climate Science Flows: Enabling Petabyte-Scale Climate Analysis with the Eart...
 
SOCRadar Research Team: Latest Activities of IntelBroker
SOCRadar Research Team: Latest Activities of IntelBrokerSOCRadar Research Team: Latest Activities of IntelBroker
SOCRadar Research Team: Latest Activities of IntelBroker
 
Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...
Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...
Multiple Your Crypto Portfolio with the Innovative Features of Advanced Crypt...
 
TROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERROR
TROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERRORTROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERROR
TROUBLESHOOTING 9 TYPES OF OUTOFMEMORYERROR
 
How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?
How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?
How Does XfilesPro Ensure Security While Sharing Documents in Salesforce?
 
Developing Distributed High-performance Computing Capabilities of an Open Sci...
Developing Distributed High-performance Computing Capabilities of an Open Sci...Developing Distributed High-performance Computing Capabilities of an Open Sci...
Developing Distributed High-performance Computing Capabilities of an Open Sci...
 

Barista: Event-centric NOS Composition Framework for SDN

  • 1. Barista An Event-centric NOS Composition Framework for Software-Defined Networks Jaehyun Nam1, Hyeonseong Jo1, Yeonkeun Kim1, Phillip Porras2, and Vinod Yegneswaran2, and Seungwon Shin1 KAIST1 and SRI International2
  • 2. Flexible Network Configurations § Software-defined networks (SDNs) • Programmable control layer à increasing network agility ・Unified integration and management of diverse functions • Deployment in various network environments ・E.g., enterprise, data center and cloud networks § Network operating systems (NOSs) • The strategic control center of SDNs ・Management of network switches and high-level applications ・Global network programmability 2
  • 3. NOS Characteristics § A variety of NOSs • Academia à Onix, Beacon, Rosemary, SE-Floodlight • Open-source projects à • Industry à § Specialized NOSs in different dimensions • Specific purposes to design NOSs, resulting in architectural differences • Examples ・Beacon for high performance ・ONOS and OpenDaylight for distributed scalability ・SE-Floodlight for enhanced security 3
  • 4. Architectural Composition Challenge § What if we need a scalable and secure NOS? • ONOS with scalability à high parallelism • SE-Floodlight with security à centralized control § Architectural Gap between NOS design principles • SE-Floodlight à ONOS ・Hard to inspect control flows due to direct comm. between components • ONOS à SE-Floodlight ・Hard to achieve high performance due to centralized inspection mechanisms ・E.g., role-based authorization, flow rule conflict resolution 4
  • 5. Barista § A novel architecture • Enable flexible and customizable instantiations of NOSs • Leverage both high parallelism and centralized control § Key Contributions • Flexible component synthesis ・Accelerate rapid prototyping and integration of new NOS functionalities • Dynamic event control ・Enable fine-grained control over events delivered to NOS components ◦ A diverse set of event types ◦ Dynamic chaining among components ◦ Policy-based event distribution 5
  • 6. Barista System Overview Operator Operator’s component 3-rd party component Barista component Barista (base framework) Barista Component pool Component management Event management Command-line interface Admin Network Network Security Security Network Management Network Base <Dynamic component event chaining> Deploy Pick-and-choose Event Worker Event Worker Event Worker <Event handler> Seq Seq Parallel Seq Seq Seq Parallel Event A Event B Event C Centralized control High parallelism 6
  • 7. Component § Implementation of a specific NOS function § Two types of components ⁃ General component ・Embedded as a part of control flows (e.g., OpenFlow engine) ⁃ Autonomous component ・Designed to take independent actions (e.g., Resource management) § Component pool ⁃ A set of components supporting the functionalities of contemporary NOSs ・Distributed scaling, performance improvement, security, failure isolation, etc. 7
  • 8. Event § Message delivery mechanism among components § No direct communication between components • All communications through the event handler § Enable high composability 8 Component #2 Component #5 Component #3 Component #n Event handler Component #1 Component #4 Event #1 Event #2 Event #3 Event #4
  • 9. Component Portability § Execute the same code of a component either inside or outside of the framework § Internal components • Higher performance • Limited flexibility § External components • High flexibility and availability • Lower performance Barista framework Source code of a component Wrapper library Source code of a component Event pipelines Control channel 9
  • 10. Diverse Event Classes § Three types of event classes • (i) Notification events, (ii) inter-component events, and (iii) meta events Event queue Worker pool C #1 C #2 C #3 Data Push Worker #1 Worker #2 Worker #n Event chain Component Pop (i) Notification events (ii) Inter-component events Look up a receiver Check event types Return results to the trigger Triggered events Reply data Push Event broker Update results Statistics Base framework Component (De)activate Collect Meta events (iii) Meta events 10
  • 11. Dynamic Component Event Chaining § Two ways to deliver events to components • Sequential vs. parallel delivery § Event sequence formulation • Ordering components based on their role and permissions • Evaluating which component can be processed in parallel or sequential Control flow integrity Flow mgmt OFP engine Rule conflict check Security / r-x Security / r-x Mgmt / r-- Base / r-- Data Role = {admin | security | network | management | base} Perm = {r (read) | w (update data) | x (cut off its control flow)} Role / Permission Control flow integrity Flow mgmt OFP engine Rule conflict check Should be in sequential Okay to be in parallel 11
  • 12. Policy-based Event Distribution § Event handling policy for deployed components § Operator-defined policy (ODP) • Datapath ID, in-port, protocol, source and dst IP addresses and ports Event queues pi pj pk … Pop Add an ODP to the target component Policy table Component Handler Check DPID Inport Proto Src IP Dst IP Sport Dport DPID1 DPID2 Any IPv4 Src IP1 Src IP2 Dst IP range 1 Any Any Skip if not matched 12
  • 13. Event Distribution Across Instances § Cluster component for event distribution • Store triggered events into a distributed storage • Keep polling new events in the distributed storage § Maintenance of logical sequence of events • Eventual consistency support § Operator-defined event sharing • Which instance listens to what kinds of events • E.g., sharing ONLY switch and topology events with other instances 13
  • 14. Implementation § Base framework and a broad set of components • 17K lines of C code and Python script • Maria DB, Galera Cluster for event distribution across instances § Open-source project • All implementation details will be provided at https://github.com/sdx4u/barista 14
  • 15. Distributed and Secure NOS § Need a distributed and secure NOS? • No controller that currently supports scalability and security together § NOS composition with the Barista component pool • Pick and choose the following components from the Barista pool ⁃ “cluster, role-based authorization, component access control, and flow rule conflict resolution” 0 200K 400K 600K 800K 1.0M 1.2M 1 2 4 8 16 32 64 128 Throughput (responses/s) The number of switches Barista ONOS SE-Floodlight 15
  • 16. Policy-based Event Distribution § Selective event distribution based on ODPs ⁃ Forwarding à DPID:!2 / DPID:2; Port:1,2; Proto:LLDP ⁃ VNF manger à DPID:2 SDN controller <Current Approach> <New Approach> Barista NOS Forwarding VNF manager Forwarding VNF manager SW#1 SW#3 SW#2 NF NF NF Unified network view Per-component network view SW#2 NF NF NF SW#1 SW#3 SW#2 Selectively deliver events Deliver all events VNF events Non-VNF events 2 1 2 1 16
  • 17. Dynamic Component Activation § Monitor event statistics in the event handler ⁃ Event volume, component level statistics regarding event consumption and production ⁃ Event distribution statistics § Dynamically trigger meta events defined by operators based on event statistics SDN controller <Current Approach> <Barista Approach> Barista NOS DDoS preventer (1-2) Run it all the time Operator DDoS attack happens! Event handler DDoS preventer (0) Define a meta event “Activate the DDoS preventer if # of new flows > 1,000 pec second” (2) Activate the DDoS preventer by itself (1) # of new flows > 1,000 per sec (1-1) Activate the DDoS preventer in manual 17
  • 18. Conclusion § Contemporary NOSs à specialized NOSs • Optimized controller design across one or a few dimensions • Fundamental differences in architectural trade-offs needed to support competing demands § Barista, A framework to build customized NOSs suited for a variety of requirements from network operators • Support flexible composition of functionalities in contemporary NOSs • Dynamic control of associated events for deployed functionalities 18 Github: https://github.com/sdx4u/barista
  • 19. Thank you for listening