Download presentation
Presentation is loading. Please wait.
1
T. Imielinski, S. Viswanathan, and B.R. Badrinath Presented by Qinhai Xia Data on Air: Organization and Access
2
Motivation Power conservation: Processor:AT&T Hobbit chip Active mode -- 25mw,Doze mode -- 50 uw Broadcasting: Quotrex system over FM channel
3
Parameter of Concern Client Channel T1 T2 T3 Tn Tuning time : T1 + T2 + T3 + … + Tn Latency: Tn - T1
4
Some Definitions Bucket bcast Index Segment Data Segment Bucket ID Bcast pointer Index pointer Bucket type
5
Latency OPT File Latency is the best: No overhead for index Latency = Data/2 + C Tuning time = Data/2 + C
6
Tuning OPT File Tuning is the best: Latency = (Data + Index) / 2 + (Data + Index) / 2 + C = Data + Index + C Tuning time = k + C k: number of levels in the index tree
7
(1, m) Indexing Data 1Data 2Data m Client Active Client Doze Tune in Next Index Pointer Client Active Client Doze Client Active Continuous Retrieving
8
(1, m) Indexing Continue Analysis: Latency = (Index + Data/m)/2 + (m*Index + Data)/2 + C = ((m+1) * Index + (1/m + 1) * Data)) / 2 + C Tuning Time = 1 + K + C K: level of the index tree
9
Distributed Indexing zNonreplicated Distribution : Different index segments are disjoint zEntire Path Replication: The path from the root to an index bucket B is replicated zPartial Path Replication (Distributed Indexing) Between two index buckets B and B’, it is enough to replicate just the path from the least common ancestor
10
Comparison Nonreplicated Distribution Full Path Replication Partial Path Replication (Distributed Indexing)
11
Distributed Indexing Analysis zr: level of index tree zLevel[r]: number of nodes on the rth level of the index tree zIndex[r]: the size of top r level of the index tree z Index r : additional index overhead Index = Level[r+1] - 1 Latency = ((Index - Index[r])/ Level[r+1] + Data/Level[r+1] + Data + Index + Index r ) /2 + C Tuning Time = 2 + k + C
12
(1, M) vs. Distributed Indexing zDistributed indexing usually has a much lower latency then (1,m) zDistributed indexing just has one more bucket overhead than (1, m)
13
Nonclustered Indexing IS Pointer to the next segment Offset to DS (K < P < L) Pointer to the next IS for value b
14
Nonclustered Indexing Protocol Client Active Client Doze Tune in Next Index Pointer Client Active Client Doze Client Active Continuous Retrieving Meta Segment Meta Segment
15
Nonclustered Indexing Analysis
16
Real Example Quotrex System Information: 160,000 bytes Bucket length: 128 bytes Bandwidth: 10 kbps
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.