A Scalable Video-On-Demand A Scalable Video-On-Demand System Using Multi-Batch System Using Multi-Batch

Buffering TechniquesBuffering Techniques

Cyrus C. Y. Choi and Mounir HamCyrus C. Y. Choi and Mounir Hamdi, Member, IEEEdi, Member, IEEE

IEEE ‘03 Transactions on Broadcasting

OutlineOutline

IntroductionIntroduction The Multi-Batch Buffer (MBB) systemThe Multi-Batch Buffer (MBB) system Normal playback requestNormal playback request VCR-like VCR-like interaction requestinteraction request Simulation resultsSimulation results

IntroductionIntroduction

The current research focus on VoD The current research focus on VoD system is on how to system is on how to lower their costlower their cost and make them and make them more scalablemore scalable..

Services providing is essential to Services providing is essential to both normal playback and both normal playback and VCR-like VCR-like interactioninteraction requests. requests.

Previous worksPrevious works

BatchingBatching Same video stream serves multiple users Same video stream serves multiple users

through through multicastmulticast Resource sharing (video stream, bandwidth)Resource sharing (video stream, bandwidth) Poor VCR-like interactionsPoor VCR-like interactions

Split and merge (SAM) Split and merge (SAM) Based on Based on batchingbatching Uses the Uses the synch buffersynch buffer to handle interaction to handle interaction

requestsrequests Demands a high I/O power in the access nodeDemands a high I/O power in the access node

Multi-Bach Buffer (MBB)Multi-Bach Buffer (MBB) ServerServer

Store video frames in the diskStore video frames in the disk Local serverLocal server

Be used to create virtual video streams (VVS)Be used to create virtual video streams (VVS) Set top box (STB)Set top box (STB)

Located at the Located at the customers’ sidecustomers’ side, can be either a hard disk or , can be either a hard disk or RAMRAM

Real video streams (RVS)Real video streams (RVS) The video streams create by The video streams create by fetching video framesfetching video frames stored stored in tin t

he diskshe disks of the video server of the video server Virtual video streams (VVS)Virtual video streams (VVS)

The video streams construct by The video streams construct by prefetchingprefetching RVSs or VVSs RVSs or VVSs at that the local servers’ buffere local servers’ buffer

STB-VVSSTB-VVS The virtual video stream created from set top box (STB) buffeThe virtual video stream created from set top box (STB) buffe

rr

Multi-Batch Buffer (MBB)Multi-Batch Buffer (MBB)

Buffer of set top box

Broadcast RVSs

Create

VVSs

Multi-Batch Buffer (MBB)Multi-Batch Buffer (MBB)

Based on a hierarchy of storage Based on a hierarchy of storage devicesdevices

Ensure Ensure multiples ofmultiples of batch time ( batch time (TTbb) ) between the start time of any two between the start time of any two video streamsvideo streams

IntroductionIntroduction The Multi-Batch Buffer (MBB) systemThe Multi-Batch Buffer (MBB) system Normal playback requestNormal playback request VCR-like interaction requestVCR-like interaction request Simulation resultsSimulation results

Handling of video requestsHandling of video requests

Reserve stateReserve state A video stream is waiting for the A video stream is waiting for the

requestsrequests Operation stateOperation state

A video stream is serving the customersA video stream is serving the customers Potential video stream (PVS)Potential video stream (PVS)

A video stream possible to be the source A video stream possible to be the source video streamvideo stream

Video start requests (VSR)Video start requests (VSR)

Case i) A PVS exists in the reserve state

Case ii) A PVS exists in the operation state

Case iii) No PVS exists request a new RVS

add to the

batch group

Video interaction requests Video interaction requests (VIR)(VIR)

Case i) Tdiff <= Tstbbuffer

Tdiff

buffering SVS

STB

Interaction stream

Interaction request

Video interaction requests (VIVideo interaction requests (VIR)R)

Case ii) Tstbbuffer < Tdiff <= Tb

STB buffer time

Tvvsdiff

buffering VVS

STB

Interaction stream

Interaction request

SVS-L

Local S

Compared with SAMCompared with SAM

Video streams always Video streams always start at i * Tbstart at i * Tb (i: integer)(i: integer) Easier to find a virtual video streamEasier to find a virtual video stream

Supplies the buffer closer to the Supplies the buffer closer to the customerscustomers Reduces bandwidth consumption of the Reduces bandwidth consumption of the

local serverlocal server

Variation of MBBVariation of MBB

MIMI 1) Improves video start requests (VSR) 1) Improves video start requests (VSR)

with VVRswith VVRs 2) Extends buffer size of STB and local 2) Extends buffer size of STB and local

server to Tserver to Tbb ( (two timestwo times the size of the size of origin)origin)

Simulation (1)Simulation (1)

Video population distributionVideo population distribution Zipf distribution: Zipf distribution: S * log (i)S * log (i)

Video interaction modelingVideo interaction modeling Probability of Probability of pp for interaction requests for interaction requests

Number of RVSs and IVSsNumber of RVSs and IVSs 50~300 and 25~200 resp.50~300 and 25~200 resp.

VSR arrival rateVSR arrival rate 60, 80, 120, 180, 360, 1200 req./hr60, 80, 120, 180, 360, 1200 req./hr

Video lengthVideo length 2 hr.2 hr.

Basic resourcesBasic resources Include RVS, VVS, STB-VVS, …Include RVS, VVS, STB-VVS, …

Result (1) Result (1)

SAM vs. MBB: arrival rate & resources requirement

S=1:

100/20 (100% of customers select the top 20% of the videos)

S=4:

80/20

S=10:

50/20

High low video probabilities distribution

Result (1)Result (1)

SAM vs. MBB:

Effective Batch time

SAM vs. MBB:

Average number of VSRs handled by a RVS

Result (2)Result (2)

Interaction handling of SAM and MBB system

Result (2)Result (2)

The reasons why the VIRs are not handled by VVS in the SAM and MBB system

Result (3)Result (3)

MBB vs. MI: arrival rate & resources requirement

Percentage of VSRs handled by VVSs in MI system with zero blocking probabilities