Virtual Active Networks Gong Su Mar. 9, 2000

Network Computing Models

  Traditional: end-to-end,  Client-server software at end nodes  The network is but a packet-transport wire Emerging: edge-to-edge    Application services/components deployed at edge nodes Examples: web proxies, firewalls, QoS/bandwidth brokers… Applications need to interact with network resources & topology   Configure resources to provide appropriate service Adapt to availability and performance of network components

VAN: Middleware for Edge-Computing

 VAN is a middleware architecture that enables applications to   Configure network topology Allocate node and link resources

A Driving Example

 A web caching application needs…  Coverage for certain network area   Connectivity among caching service components Resources to move cached contents  Solution: requests a VAN that provides   Coverage: spans a ring between AS1, AS2, AS4, and AS5 Resources: provides at least 1mbps for all connections  1mbps Reliability: prohibits more than 2 virtual links from traversing the same physical link Physical network Virtual spec.

AS1 D 1mbps A B C AS2 1mbps AS3 D 1mbps C Logical hierarchy A D B C Mapping by VAN AS5 B A AS4

VAN Contributions

   Enable applications to configure network  Algorithm that maps VAN specification to physical resources Acquire distributed node and link resources  Deadlock-free VAN resource provisioning protocol Recover from underlying network failure  Protocol that preserves VAN service semantics under failures

VAN Service Arch Components

 

VAN Local Manager

 (VLM) Manages local node resources   Supports deadlock-free VAN provisioning Monitors & reports resource status

VAN Domain Server

  (VDS) Provides VAN services to application VAN provisioning   Resource acquisition Performance monitoring  Manages VAN to recover from physical network failure Active node with VLM AS3 AS1 D C AS2 VDS VDS AS5 B A AS4 VDS administrative domain

Specification Mapping



Heuristic mapping algorithm

 Sort VNs and PNs by degree; map VN to PN by degree-order   Mark all PL without enough bandwidth for the VLs as infeasible Each PL has a “mapped-onto” counter, initially 0    pick a VL and map it to a physical path with lowest maximum counter among all PLs traversed After each VL is mapped, increment counter and subtract available bandwidth for each PL; mark a PL infeasible as appropriate Repeat until all VLs are mapped AS3 AS1 D 1mbps A 1mbps B 1mbps D 1mbps C C AS2 A AS4 AS5 B

Resource Acquisition Protocol

   Acquires node and link resources   Intra-domain: VDS – VLM Inter-domain: VDS – VLM and VDS – VDS Deadlock among competing VANs for shared resources can occur because   One VAN is built in many domains distributedly Many VANs are built in many domains simultaneously Example     VDS1 and VDS2 build VAN1 and VAN2 in domain A respectively VDS3 and VDS4 build VAN1 and VAN2 in domain B respectively VAN1 preempts VAN2 in domain A VAN2 preempts VAN1 in domain B B A VDS4 VDS2 VDS3 VDS1 VAN1 VAN2 VAN1 VAN2 VLM2 VLM1

Deadlock Prevention Protocol

 1 How does the solution work  Assign “weight” to VNs and VLs   Each VDS computes a “Progress Index” (PI), indicating “how much” a VAN has been built PIs are globally synchronized and used as the priority for preemption when conflict 2 VLM detects conflict and initiates arbitration 3 VDS broadcasts to all other VDS’es requesting global PI 4 5 VDS’es ack arbitration request 6 VLM notifies VDS’es with the arbitration decision A VDS1

VAN1

:3 VLM1 VDS2

VAN2

:6 A VDS1 1 2 VLM1 2 1 VDS2 4 3 3 4 A VDS1 5 6 VLM1 6 5 VDS2 B VDS4 VDS3

VAN1

:7 VLM2

VAN2

:2 VDS3 VDS3 B VLM2 B VLM2 VDS4 VDS4

Failure Recovery

   When a physical link fails, the VLs it carries must be restored First try Local repair    Find an alternative path with adequate resources between the two disconnected AS’es Fast, and preserve original topology But   May violate reliability constraint Alternative path may not exist Example  Physical link between 2 and 3 goes down  Alternative path goes through 2, 1, 4, and 3 2 Physical link Virtual link 1 5 4 6 2 3 Reliability violation 1 5 4 6 3

Failure Recovery: Global Repair

  Local repair may violate reliability constraints or it may not be able to find an alternative path Global repair    Computes substitute VL based on global topology and resource information Reconstruct topology when local repair cannot; guarantee reliability constraint But   Computationally expensive Communication delay between root VDS and local VDS’es  Example  Substitute VL computed between 5 and 6, replacing the VL going VDS1 2 2 1 1 Physical link Virtual link 4 3 4 3 VDS3 VDS2 5 6 VDS3 VDS2 5 6

Schedule

End of Spring 2000 Summer 2000 (first half) Summer 2000 (second half) Fall 2000 (first half) Fall 2000 (second half) Efficiently obtain global topology and resource information Heuristics for virtual specification to physical network mapping with constraints Algorithm for computing dynamic priority (PI) and analyze conflict resolution protocol Efficient local repair mechanism (study MPLS fast rerouting, ATM self-healing, etc.) Incremental global repair mechanism