1. Distributed Dynamic Replica Placement and Request Redirection in Content Delivery Networks Advisor : Ho-Ting Wu Student : Yu-Chiang Lin Date:2011/5/30

2.  Page 2 OUTLINE  CDN introduce  Distributed Dynamic Replica Placement  Cloning of a replica  Replica removal  Redirection  Future work  Reference

3.  Page 3 OUTLINE  CDN introduce  Distributed Dynamic Replica Placement  Cloning of a replica  Replica removal  Redirection  Future work  Reference

4.  Page 4 CDN Introduce Fig:Server Farm(source from EBU technical review)

5.  Page 5 CDN Introduce  Consider TCP transmittion ， throughput may affect by latency or packet lost 。

6.  Page 6 CDN Introduce  The Content Delivery Networks (CDN) paradigm is based on the idea to transparently move third-party content closer to the users  content is replicated on CDN servers which are located close to the final users so that users perceive a better content access service.

7.  Page 7 CDN Introduce Fig:Content Delivery Network(source from EBU technical review)

8.  Page 8 CDN Introduce Fig:Content Delivery Network

9.  Page 9 CDN Introduce  Four Important technique  1.Content route  2.Content distribution  3.Content store  4.Content management

10.  Page 10 CDN Introduce  CDN issue  1) Deciding the kind of content that should be hosted (replica placement)  2) selecting the best replica for a given user  3) designing mechanisms for transparent redirection of the users requests to the best replicas

11.  Page 11 OUTLINE  CDN introduce  Distributed Dynamic Replica Placement  Cloning of a replica  Replica removal  Redirection  Future work  Reference

12.  Page 12 Distributed Dynamic Replica  Distributed scheme to allocate and deallocate replicas, so that the user requests are satisfied while minimizing the CDN costs in a dynamic scenario  This scheme always accounts for the current replica placement,adding replicas or changing replica location only when needed.  Each site j ∈ V R autonomously decides on whether some of the replicas it stores should be cloned or removed.

13.  Page 13 Problem Formulation  V A : access nodes  V R : CDN servers sites  d(i, j): user (access node) i to a replica j  d max : distance threshold  x i,c : the volume of user requests originated at node i ∈ V A for content c ∈ C  U max, U mid, U low: load threshold

14.  Page 14 Cloning of a replica  Function add_replica(j, c)  1: l j,c =i ∈ V A α ij,c · x i,c  2: while l j,c /r j,c − U max > 0  3: best_served = 0  4: best_distance = ∞  5: best_vr = undefined  6: for all j’ ∈ ρ(j) s.t. r j,c < V maxR do

15.  Page 15 Cloning of a replica  7: l’ j,c =i ∈ α(j’) α ij,c · x i,c  8: total_distance =i ∈ α(j’)α ij,c · x i,c · d i,j’  9: if (l’ j’,c < best_served) ∨  10: (l’ j’,c = best_served ∧  11: total_distance < best_distance) then  12: best_distance = total_distance  13: best_served = l’ j’,c

16.  Page 16 Cloning of a replica  14: best_vr = j’  15: end if  16: end for  17: if best_vr = undefined then  18: exit  19: end if  20: ask best_vr to add a replica

17.  Page 17 Cloning of a replica  21: compute l’’ best vr,c = min(i ∈ α(best vr ) α ij,c · x i,c, 1)  22: l j,c = l j,c − l’’ best vr,c  23: remove from the set of requests those that can be offloaded  24: end while

18.  Page 18 Replica removal  A replica can be removed if (and only if) it serves no requests  A replica is removed only when it has not been serving requests for a time long enough to bring the exponential average down to zero.

19.  Page 19 Redirection  Based on this feedback users requests are directed away from a replica in response to threshold events (if the replica load exceeds U max or falls below U low )  As an example, an underloaded replica informs the redirection system which then tries to offload requests to some other replicas (if possible)  A perfect load balancing may be impossible due to the distance constraint

20.  Page 20 Redirection Fig. A model for the redirection scheme

21.  Page 21 OUTLINE  CDN introduce  Distributed Dynamic Replica Placement  Cloning of a replica  Replica removal  Redirection  Future work  Reference

22.  Page 22 Future Work  comparison of different settings of the U mid  observe d max and add remove replica relationship

23.  Page 23 Reference  [1] N. Bartolini, F. Lo Presti, and C. Petrioli, “Optimal dynamic replica placement in Content Delivery Networks,” in Proceedings of the 11 th IEEE International Conference on Networks, ICON 2003, Sydney, Australia, September 28– October 1 2003, pp. 125–130.  [2] F. Lo Presti, C. Petrioli, and C. Vicari, “Dynamic replica placement in content delivery networks,” in Proceedings of MASCOTS 05, September 2005.  [3] F. Lo Presti, C. Petrioli, and C. Vicari, “Distributed Dynamic Replica Placement and Request Redirection in Content Delivery Networks,” in Proceedings of MASCOTS 07, 2008.