Open Programmable Architecture for Java-enabled Network Devices Tal - PowerPoint PPT Presentation

Open Programmable Architecture for Java-enabled Network Devices Tal Lavian Technology Center Nortel Networks tlavian@NortelNetworks.com Santa Clara University 9/29/99 1

Programmable Network Devices Openly Programmable devices enable new types of intelligence on the network Santa Clara University 9/29/99 2

Agenda • Local Computation • New types of applications • Architecture • API’s • Summary Santa Clara University 9/29/99 3

Changing the Rules of the Game • Move Turing Machine onto device — Add local intelligence to network devices while (true) { doLocalProcessingOnDevice() } Santa Clara University 9/29/99 4

Technology Concept Applet Web Browser Web Server The JVM is in the Browser Download applications for local processing Download applications for local processing non-bundled application Server Reversed Applet Santa Clara University 9/29/99 5

The Web Changed Everything • Browsers — Introducing JVM to browsers allowed dynamic loading of Java Applets to end stations • Routers — Introducing JVM to routers allows dynamic loading of Java Oplets to routers This Capability WILL Change Everything Santa Clara University 9/29/99 6

Accomplishments —JVM on a silicon-based Routing Switch —ORE - Oplet Run-time Environment —Java-enabled Device Architecture —Java SNMP MIB API —Implementation of Network Forwarding API —All of this enables implementation of Dynamic Classification in Silicon-Based Forwarding Santa Clara University 9/29/99 7

Paradigm Shift • Supports distributed computing applications Java-based in which network Application devices participate Java-based Application — router to router — server to router • Supports Intelligent Agents Java-based Application • Supports Mobile Agents Santa Clara University 9/29/99 8

Example: Downloading Intelligence Example: Downloading Intelligence Monitor Authentication Security Intelligence Dynamic application React loading JVM OS HW Network Device Santa Clara University 9/29/99 9

Security and Stability • secure download of Java Applications • safe execution environment — insulate core router applications from dynamically loaded applications Santa Clara University 9/29/99 10

Device-based Intelligence • Static-vs-Dynamic Agents — Static – SNMP set/get mechanisms – Telnet, User Interfaces (cli, web, etc…) — Dynamic closed-loop interaction on nodes – capable of dealing with new and difficult situations – autonomous and rational properties. – dynamically system monitoring & modification – report status and trends Santa Clara University 9/29/99 11

Agenda • Local Computation • New types of applications • Architecture • API’s • Summary Santa Clara University 9/29/99 12

New Types of Applications • Mobile Agents • Local Intelligence for NMS • Collaboration among routers • Router & Server Collaboration • E-commerce Santa Clara University 9/29/99 13

Mobile Agents • Intrusion Detection - Hacker Chaser • Trace-route for Layer 2 • Mobile Connectivity Mapper Santa Clara University 9/29/99 14

Local Intelligence for NMS: Diagnostic Agents No more • Download Intelligent Agent Extensive access polling to internal resources monitor from NMS to the device. Download • Wait for threshold. Monitor • Might be complex conditions Complex Condition • Trend analysis Exceeded • Send “condition exceeded” NMS event to NMS. • Automatic download appropriate application Appropriate • Application takes action. Download Application router Santa Clara University 9/29/99 15

Application Layer Collaboration Among Routers and Servers • Application aware routing • Server farm load balancing — server state monitored — rerouting based on congestion/load • Auctioning Applications Santa Clara University 9/29/99 16

Applications Aware Forwarding Business logic based operation changes • Resize forwarding queues • Modify congestion control algorithm • Adjust Packet Scheduling • Change routing table Santa Clara University 9/29/99 17

Agenda • Local Computation • New type of applications • Architecture • API’s • Summary Santa Clara University 9/29/99 18

ORE - Oplet Run-time Environment Oplet 1 Oplet 2 Service A Service B Service C Why ORE? ORE JVM Santa Clara University 9/29/99 19

Node Architecture Node Architecture Download Oplet C/C++ Java API API ORE Service JNI Device Oplet Runtime Env Code I P A Device JVM D Drivers W F Operating System J Device HW Santa Clara University 9/29/99 20

Architecture Issues • Green Threads -vs- Native Threads — Native threads: – provides non-interference between Java applications – difficult thread-to-thread communication and sharing of data between threads – creates a dependency on underlying RTOS – multiple JVM instances consume resources — Green Threads – single JVM must manage CPU & memory resources between concurrently running threads Santa Clara University 9/29/99 21

Evolution of Router Architecture Distributed, Distributed, Centralized, Centralized, line-card based Model line-card based Model CPU-based Model CPU-based Model Routing software Routing software Routing software Routing software (forwarding (forwarding buffering) w/ router OS w/ router OS Line card buffering) w/ router OS w/ router OS Line card Routing software Routing software Routing software Routing software w/ COTS OS w/ COTS OS w/ COTS OS w/ COTS OS Routing Buffer Routing Buffer memory memory CPU CPU Buffer Buffer CPU CPU memory memory Line card Line card Line card Line card (forwarding (forwarding (forwarding (forwarding buffering) buffering) buffering) buffering) line card line card line card line card line card line card ... ... NI as NI as NI as NI as NI as NI as (forwarding (forwarding buffering) Line card buffering) Line card Added scalability, Flexibility, extensibility Control + Forwarding Control + Forwarding Control separated Control separated Functions combined Functions combined From forwarding From forwarding Santa Clara University 9/29/99 22

Explicit Separation of Control Plane from Data Forwarding Control Element g n i t u o R Shared Traditional device Forwarding Memory Forwarding/ Flow/filter Forwarding Table Element Download CPU Forwarding Packet Line Card Flow Forwarding Element Line Card Forwarding Forwarding Element Forwarding Santa Clara University 9/29/99 23

Separation of Control and Forwarding Planes Centralized, Centralized, Forwarding-Processors Forwarding-Processors CPU-based Router CPU-based Router based Router based Router Routing Control SW Plane CPU CPU Forwarding Forwarding Forwarding Processor Processor Processor Slow Wire Speed Control + Forwarding Control + Forwarding Control separated Control separated Functions combined Functions combined From forwarding From forwarding Santa Clara University 9/29/99 24

Open Networking Architecture U n i f i e d IP Telephony Policy Server VPN Firewall p o l i c y - b a s e d Application Vertical Proprietary Server Operating System server Networking Box Level Network Services Protocol Hardware Proprietary Apps Control Network Services Objects element Network OS Proprietary NOS Connect Transport Interface Custom Switch ASIC’s ma n a g e me n t Forwarding Real-time OS Today element Network Si Open Santa Clara University 9/29/99 25

Dynamic Configuration of Forwarding Rules Dynamic Configuration of Forwarding Rules AN Apps CPU Forwarding Forwarding Forwarding Forwarding Rules Rules Rules Rules Forwarding Forwarding Forwarding Forwarding Processor Processor Processor Processor SW HW Santa Clara University 9/29/99 26

Real-time forwarding Stats and Monitors Real-time forwarding Stats and Monitors AN Apps CPU Forwarding Forwarding Forwarding Rules Rules Rules Forwarding Forwarding Forwarding Processor Processor Processor Statistics Statistics Statistics &Monitors &Monitors &Monitors SW HW Santa Clara University 9/29/99 27

Dynamic - On the Fly Configuration Dynamic - On the Fly Configuration Policy AN Apps Filters Filter Packet Packet Forwarding Forwarding Processor Processor Packet Santa Clara University 9/29/99 28

Active Networks Packet Capture Active Networks Packet Capture AN Apps JFWD to Divert or Copy CPU Wire Speed Forwarding Forwarding Forwarding Forwarding Processor Processor Processor Processor Packet Santa Clara University 9/29/99 29

Scaling up Active Networks Routing Protocol to commercial networks • Overcome the need to predefine the next hop Overcome the need to predefine the next hop • No need to know AN topology a head of time • Divert/CarbonCopy specific packets to control plane (e.g. packets on ANEP port ) • Wire speed of all other packets • End to end forwarding • Future: Active Networks Routing Protocols Santa Clara University 9/29/99 30