1 / 25

Resource Management with Virtual Paths in ATM Networks

Resource Management with Virtual Paths in ATM Networks. Wanru Tao SUNY at Buffalo. Topics of Discussion. Backgrounds (VP, VC, VPC networks, VPC Provisiong etc) Overview Discuss Four Distinct VPC management - Successive Capacity Reallocation - Successive Topology Reconfiguration

ivie
Télécharger la présentation

Resource Management with Virtual Paths in ATM Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Resource Management with Virtual Pathsin ATM Networks Wanru Tao SUNY at Buffalo

  2. Topics of Discussion • Backgrounds (VP, VC, VPC networks, VPC Provisiong etc) • Overview • Discuss Four Distinct VPC management - Successive Capacity Reallocation - Successive Topology Reconfiguration - Global Reconfiguration - Long - Term planning

  3. VCI and VPI translation VCC VPC1 VPC2 A B c E D F VPI =4 VPI=15 VPI=8 VPI=10 VPI=25 VCI=6 VCI = 6 VCI=6 VCI=6 VCI=6 VPC terminator Transit node

  4. Overview • ATM Network is organized as a collection of VPCs which form a VPC, or logical, overlay network • If the VPCs are permanent or semi-permanent and have reserved capacity, establishing new VCCs requires simple connection admission decisions at the VPC terminators of existing VPCs(faster connection) • To which provisioning is able to improve network efficiency is dependent on the resource management decisions that determine the VPC topology and capacity allocation • VPC Provisioning improve network efficiency depends on - Its ability to provide VCCs with low setup and switchingcosts, while maintaining a low call blocking probability( in turn, is dependent on the VPC topology and capacity allocations resulting from resource management decisions

  5. VPC Provisioning improve network efficiency depends on(cont’d) - switching cost increases with increase of physical hop count, the number of VPC. - if VPC route is available with sufficient capacity to accept the new VCC, setup cost is relatively low, since decisions are made exclusively within the control plane(i.e.,by CAC and routing algorithms) - if a VPC route exists, but does not have adequate capacity to accommodate the new VCC, then the only way to avoid call blocking is to allocate additional capacity to some set of VPCs - if no VPC route is found from source to destination, then the only way to void call blocking is by setting up a new VPC or set of VPCs • The above two scenarios incur higher cost

  6. View the provisioning of resources in a VP-based network in three time scales -resource management which modify the VPC topology and capacity allocations in order to adapt to changing traffic conditions and possible network failures - Such adaptation can be categorized by the timescale on which they operate • Long term - corresponding to the physical network • medium-term - VPC overlay network • short-term - individual VCCs On the timescale of VPC provisioning, resource management activities are more likely to be handled by the management plane and include VPC topology and VPC capacity allocation decisions

  7. Four Distinct VPC management are discussed • successive capacity reallocation - redistributes capacity on a fixed VPC topology - concerned with changes in capacities allocated to specific VPCs • successive topology reconfiguration - This involves a slight modification to the VPC overlay network (e.g, the addition of a new VPC to facilitate resource management). - established and /or tears down VPCs within an existing VPC topology

  8. Four Distinct VPC management are discussed (Cont’d) • Global reconfiguration consists of both global capacity reallocation and global topology reconfiguration. This activity potentially affects all VPCs in the network - significant changes to the VPC topology, and corresponding capacity allocations are performed • Long-term planning derives a static(or general) set of VPCs and initial or minimum capacity assignment of them. - Is applicable to VPC topological design at startup or after a physical network upgrade.

  9. Each Discuss will around three aspects of VP resource management • triggering mechanism - concerned with conditions under which changes are to be made to the VPC topology and /or capacity allocations • Design algorithm - address the question of what changes should be made • implementation - concerned with how the changes should be put into place

  10. Successive Capacity Reallocation Triggering Mechanism • On demand - Ohta and Sato said, to trigger capacity increase on a set of VPCs whenever a new connection cannot be accommodated - Reduction of capacity can also be initialized on demand by diminish the extra capacity which has finished finishing using • When some threshold of spare capacity has been reached, additional capacity can be request. - This method attempts to react to traffic fluctuations before blocking occurs, with advantage that connection setup need not be be delayed • When performance monitoring indicated a need for reallocation - when performance degrades beyond some state, the capacity reallocation alg. is invoked • When failures require transfer of capacity

  11. Successive Capacity Reallocation ( Cont’d) Design Algorithm • Goal of a Successive reallocation is either to reduce the call blocking probability or ensure the efficient use of resources. Other factors include responsiveness, stability, and fairness • changes to the capacity of a VPC can be made in quanta(add/delete in step sizes). To achieve fairness in the sense of equal blocking, large step size for heavily used VPCs and a small step size for lightly loaded VPCs. • How to acquire the additional capacity for a capacity increase • if there is available spare capacity, the reassignment is straightforward • VPCs with lowest call blocking probability will be chosen for capacity reduction

  12. Successive Capacity Reallocation (Cont’d) Implementation • By changing bandwidth tables used by the CAC algorithms. • capacity updates are implemented in two stages - logical bandwidth modification makes changes to resource that are perceived as available to the CAC algorithms ; this modification is done first so that the occupied capacity can be decreased as the calls are terminated - physical bandwidth modification concerned with changes to policing and shaping parameters, as well as buffer sizes assigned to the VPC when the occupied capacity reached the target level, physical bandwidth modification can be done When the physical bandwidth modification is completed. The released bandwidth becomes available for reallocation

  13. Successive Topology Reconfiguration Triggering Mechanism & Design Algorithms(tend to be closely related) • Goal of successive topology reconfiguration is to decrease the amount of processing at transit nodes • accomplished by providing shorter paths(I.e..., direct routes) for source-destination pairs with high traffic demand • A centralized network management facility responds to network congestion by identifying congested pairs of nodes and initiating a topology reconfiguration • reduce the overall VPC hop count, which had a direct affect on VCC setup

  14. Successive Topology Reconfiguration Triggering Mechanism & Design Algorithms (Cont'd) • Usage patterns are monitored in order to learn which end-to-end paths become more heavily used so that the number of VPC hops can be decreased for these paths • At given intervals, the VPC layout is reassessed to determine whether or not VPC additions and deletions would be beneficial.

  15. Successive Topology ReconfigurationImplementation • 1st phase in implementing a topology change is - To place the new VPCs in the affected routing tables - A VPC setup policy then decides when these VPCs should become usable • 2nd phase of implementation - Either wait for the Vccs on the old VPCs to terminate or reroute Vccs from the old VPCs onto the new VPCs • 3rd Phase of implementation Is to tear down the old VPCs

  16. Successive Topology ReconfigurationImplementation • Gerla stressing the need for transparency, suggest a gradual movement from the old to the new VPCs - A small amount of capacity can be reserved for the new VPCs in order to accommodate new VCCs - Old VPCs are torn down as connections terminate, allowing for more capacity to be transferred to the new VPCs

  17. Global Reconfiguration • global reconfigurations useful when modifications are periodically made based on anticipated traffic conditions(e.g., hourly or daily). • It is also useful in facilitating recovery from component failure • Given a specification of the expected traffic, most solutions presented for global modifications calculate all of VPC topology, VPC capacity allocations, and VCC routing. • The resulting problem is typically formulated as an optimization of some kind

  18. Global ReconfigurationTrigger Mechanism • A possible trigger is based on traffic prediction, which is used as a guideline to altering the VPC overlay network. • observation degradation in performance caused by traffic variations can also trigger VPC reconfiguration. • Logothetis and Shioda propose an activity where global capacity reallocations are made when the observed call blocking probability exceeds some threshold.

  19. Global ReconfigurationDesign Algorithms Since global reconfiguration involves both topology and capacity, our approach is to first discuss the issues associated with VPC topology design and capacity allocations, and then discuss the solutions to the various optimization problems.

  20. Design Algorithms • global reconfiguration involves both topology and capacity. - VPC topology design - Capacity allocation

  21. Global ReconfigurationTopology Design • In general, the number of possible VPC topologies grows exponentially with the network size • The issue of how large a change can be made to the VPC overly network during normal network operation without causing service disruptions is not well understood

  22. Topology Design • To reduce the complexity of the design problem, only a subset of the possible paths are considered for VPC setup. • Some studies only consider direct VPCs between source-destination pairs with VPC hop count at most two • trade-off between a low setup cost, which favors a low VPC hop count, and efficient resource usage, which favors a higher VPC hop count.

  23. Capacity Allocation • focuses on performance( minimizing the VCC blocking rate or maximizing the carried load) • Fairness. If traffic is segregated by Cos(class of service), this segregation can be used to enforce fairness by controlling the amount of capacity allocated to each Cos

  24. Solution • Solution by Gerla, finds the physical routes and capacities of VPCs, such that the average delay is minimized. The algorithms limits the candidate VPCs so that direct to VPCs are placed between heavily used pairs of nodes, which indirect VPC routes are used to connect less loaded pairs • LEE and YEE purpose optimization problem to minimize the total cell loss rate and find VPC routes along with capacity assignment • Hui purpose optimization to ensure low probability of blocking a connection on a direct route from src to dest.

  25. End of presentation

More Related