1. Placement of Continuous Media in Wireless Peer-to-Peer Network Shahramram Ghandeharizadeh, Bhaskar Krishnamachari, and Shanshan Song IEEE Transactions on Multimedia, April 2004

2. H2O Framework  Home-to-Home Online (H2O) devices collaborate to deliver continuous media  H2O may act as: A producer of data An active client A router

3. Motivation  A new replication technique that Provide on-demand access to continuous media Minimize the total storage space required

4. Assumptions  CBR continuous data  Total size of available clips exceeds the storage capacity of one device  Bandwidth between two H2O devices exceeds the bandwidth required to display a clip  One hop distance is a constant

5. H i : the Farthest Number of Hops a Block Can be Located  Cycle: period to display a block D=S b /B Display  The farthest number of hops that the block i can be located: H i =((i-1)D)/h block sizeplayback rate time to retrieve a block from one hop away

6. Data Placement and Replication  For each video clip X: Divide X into equal-sized blocks with size S b Place first block, b 1 on each node. For each block b i, 1<i<=z, compute delay tolerance H i Compute r i based on H i Construct r i replicas of b i and place them  r i is a topology dependent computation

7. Topology I: Worst Case Linear Topology  Block i should be replicated r i times: H i =(i-1)D/h r i =N-H i Reset r i to one if r i is zero or negative  Total storage space (S C,R ) occupied by a clip with z blocks: 12389 …

8. Percentage Saving Compared with Full Replication in Linear Topology N=1000, h=0.5, B Display = 4Mbps y: 100x(1-S C,R )/(S C xN)

9. Topology II: Grid Topology  Organize N nodes in a square area  At least one copy of b i must be placed within H i hops There are nodes within H i hops of every node  Total storage required:

10. Total Storage Space Required as a Function of Block Size (1/2) h=0.75s 2 min clip (total 60MB)

11. Total Storage Space Required as a Function of Block Size (2/2) h=0.75s 2 hour clip (total 3600MB)

12. Topology III: Average Case Topology (1/2)  Network connectivity depends on radio range R  N nodes are scattered in area A  There are on average between and nodes within H i nodes.

13. Topology III: Average Case Topology (2/2)  Using the upper boundary, the H number of replicas r i required by b i is:  Total storage required for a clip:S

14. Percentage Saving Comparison

15. Distributed Implementation  H2O p : publish a clip X Compute block size S b, number of blocks z, and H i for each block Flood the network to query which H2O will host a copy of which block of X  H2O j : each recipient of the message Compute a binary array A j that consists of z elements whose values are 0 or 1 Two computation methods: TIMER or ZONE

16. Technique I: TIMER  When H2O j receives query message Perform z rounds of elections Pick a random timer value between 1 and M then count down The one first count down to zero stores a copy and send suppress message within H i hops  May generate more than one copies of a block within H i hops

17. Technique II: ZONE  Assume each node is aware of its (x, y) coordinate  Place each copy in a separate square zone whose size is such that all nodes can be reached within H i hops

18. Simulation: TIMER vs. ZONE N=300, R=100m, A=1km2, z=60

19. Simulation: Comparison of Analytical Models for Graph Topology with 2 Implementations  S C =60MB  R=100m  A=1km2

20. Simulation: How Many Blocks a H2O Device Have When Using TIMER N=300, R=100m, A=1km2 Average # of blocks per node for a clip is marked as dashed line

21. Conclusion  Provide a novel replication technique for on-demand clips Minimize startup delay Storage saving compared with full replication  Provide two distributed implementations

22. How Many Nodes Within H i Hops in Grid Topology Take H i = 3 for example: (3+1)*(3+1) 3*3 (H i +1)*(H i +1)+H i *H i =

23. How Many Nodes Within H i Hops in Average Case Topology R … HiRHiR The circle area with radius H i R= / A ≒ # of neighbors / N # of neighbors = Density dependent variable