1 / 22

Dynamic RWA

Dynamic RWA. Connection requests arrive sequentially. Setup a lightpath when a connection request arrives and teardown the lightpath when a connection departs Goal is to minimize connection blocking Solve the routing subproblem and the wavelength assignment subproblem separately. Routing.

cliff
Télécharger la présentation

Dynamic RWA

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. Dynamic RWA • Connection requests arrive sequentially. • Setup a lightpath when a connection request arrives and teardown the lightpath when a connection departs • Goal is to minimize connection blocking • Solve the routing subproblem and the wavelength assignment subproblem separately

  2. Routing • Fixed routing • Fixed-alternate routing • Adaptive routing

  3. Fixed Routing • Always choose the same fixed route (calculated offline) for a given source-destination pair • E.g. shortest-path routing • Advantage: simple • Disadvantages: • High connection blocking • Unable to handle faults

  4. Fixed-Alternate Routing • Each node maintains an ordered list of a fixed set of routes to each destination node • E.g., k shortest-path routes • Primary route: the first route in the list • Alternate route: a route that does not share any links with the first route • Useful for fault tolerance

  5. Fixed-Alternate Routing • When a connection request arrives, the source node tries each of the routes in the list in sequence until a route with a valid wavelength assignment is found • Advantages • Simple • Fault tolerance • Significantly reduce the connection blocking probability compared to fixed routing

  6. Adaptive Routing • Route is chosen based on the current network state • Two approaches • Adaptive shortest-cost-path routing • Least-congested-path routing

  7. Adaptive Shortest-Cost-Path Routing • Use layered graph • Link costs • 1 for unused link •  for used link • c for wavelength conversion link • When a connection request arrives, compute the shortest-cost path between source and destination • Advantage: low blocking • Disadvantage: nodes need update network state whenever a connection is setup/teardown

  8. Least-Congested-Path (LCP) Routing • For each s-d pair, a set of routes is predetermined • When a connection request arrives, the least-congested path is chosen • Congestion on a link = # wavelengths available on the link • Fewer available wavelength  more congested • Congestion on a path = congestion on the most congested link in the path • Use shortest-path routing to break ties • An alternative: give priority to shortest paths, use LCP to break ties

  9. Wavelength Assignment Heuristics • Assume fixed number of wavelengths • Minimize overall blocking probability for all connection requests • Single-fiber networks: R, FF, LU, MU • Multi-fiber networks: MP, LL, M, RCL • Protect multihop connections to achieve greater degree of fairness • Rsv, Thr

  10. Wavelength Assignment Heuristics: Single-Fiber Case • Random Wavelength Assignment (R): • Find all wavelengths available on the required route • Randomly choose one available wavelength • First-Fit (FF) • Wavelengths are numbered • Choose the first available wavelength • Computation cost lower than R • Perform well in terms of blocking probability and fairness • Both R and FF require no global knowledge

  11. Wavelength Assignment Heuristics: Single-Fiber Case • Least-Used (LU)/SPREAD • Choose the least used wavelength • Attempt to balance the load among all wavelengths • Favor short paths, not fair for long paths • Perform worse than random • Most-Used (MU)/PACK • Choose the most used wavelength • Pack connections into fewer wavelengths • Slightly outperform FF • Both LU and MU require global knowledge

  12. Wavelength Assignment Heuristics: Multi-Fiber Case • Min-Product (MP) • Goal: minimize # fibers by packing wavelengths into fibers • First compute for each wavelength j that is available on p • Choose the lowest numbered wavelength in the set of wavelengths that minimize the above value • Become FF in single-fiber networks • Perform worse than the multi-fiber version of FF (both fibers and wavelengths are ordered)

  13. Wavelength Assignment Heuristics: Multi-Fiber Case • Least-Loaded (LL) • Select the wavelength that has the largest residue capacity on the most loaded link along route p • Choose the minimum indexed wavelength j in Sp that achieves • Become FF in single-fiber networks • Outperform MU and FF

  14. Wavelength Assignment Heuristics: Multi-Fiber Case • MAX-SUM (M) • Assume the set of possible connection requests is known in advance and the route for each connection is pre-selected • Attempt to maximize the remaining path capacities after lightpath establishment

  15. Wavelength Assignment Heuristics: Multi-Fiber Case • MAX-SUM (M) • : a network state that specifies the routes and wavelength assignments of existing lightpaths • Link capacity r(, l , j) on link l and wavelength j in state : # fibers on which wavelength j is unused on link l • Path capacity r(, p , j) on path p and wavelength j: # fibers on which wavelength j is available on the most congested link along path p • Path capacity of path p in state , R(, p): sum of path capacities on all wavelengths

  16. Wavelength Assignment Heuristics: Multi-Fiber Case • MAX-SUM (M) • ’(j): the next state of the network if j is assigned to the connection • P: set of all potential paths for connection requests in the current state • Choose the wavelength j that maximizes • Equivalently, choose wavelength j that minimizes the total capacity loss on this wavelength, which is

  17. Wavelength Assignment Heuristics: Multi-Fiber Case • Relative Capacity Loss (RCL) • Improve on M by taking into consideration # available alternate wavelengths for each potential future connection • RCL chooses wavelength j that minimizes the sum of the relative capacity loss on all the paths

  18. Heuristics for Protecting Multihop Paths • Longer lightpaths have a higher probability of getting blocked than shorter paths  want protect longer paths • Proposed schemes: Rsv and Thr • Only specify whether the connection request can be assigned a wavelength under the current wavelength-usage conditions  must be combined with other wavelength assignment schemes • Achieve a greater degree of fairness

  19. Heuristics for Protecting Multihop Paths • Wavelength Reservation (Rsv) • A given wavelength on a specified link is reserved for a multihop traffic stream • Reduce blocking for multihop traffic while increasing the blocking for single-hop traffic • Protecting Threshold (Thr) • A single-hop connection is assigned a wavelength only if the number of idle wavelengths on the link is at or above a given threshold.

More Related