Dissemination of Dynamic Data and Information Services on Broadcast Channels
博士 === 國立臺灣大學 === 電機工程學研究所 === 91 === The needs of various classes of modern data and information applications, especially those targeted at mobile computing, demand the support of adaptive data broadcasting, and the support in turn demands a new kind of d...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2003
|
Online Access: | http://ndltd.ncl.edu.tw/handle/34319669630449783038 |
id |
ndltd-TW-091NTU00442144 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-091NTU004421442016-06-20T04:15:46Z http://ndltd.ncl.edu.tw/handle/34319669630449783038 Dissemination of Dynamic Data and Information Services on Broadcast Channels 廣播頻道上動態資料與資訊服務傳播 Chih-Lin Hu 胡誌麟 博士 國立臺灣大學 電機工程學研究所 91 The needs of various classes of modern data and information applications, especially those targeted at mobile computing, demand the support of adaptive data broadcasting, and the support in turn demands a new kind of data dissemination framework. For proper implementation, the service model of adaptive data broadcasting must provide the methodologies of dynamic traffic awareness and dynamic broadcast content adaptation. At the same time, these methodologies must be implemented as a distributed data dissemination platform, having the guarantees of efficiency, reliability, robustness, and scalability. We advocate that the design and development of adaptive data broadcasting schemes should be backward compatible, rather than replacing the conventional paradigm, in accordance with the principles of information dissemination and dynamic data management. To this end, in this dissertation, we formulate the fundamentals of adaptive data broadcasting, relate them to the traditional data broadcast scheme, and devise several promising mechanisms. Note that prior researches in the data broadcast realm are mainly based on the traditional data management model, where data items are independent, persistent, and static. In fact, however, many modern information broadcast applications enable the generation and spread of dynamic data. To fulfil the requirements for the dissemination of dynamic data and information broadcast services on broadcast channels, we address the impacts of dynamic traffic awareness and dynamic broadcast content adaptation. Accordingly, we propose in this dissertation four promising mechanisms for this purpose:\ particularly, (1) Dynamic Broadcast Traffic Awareness, (2) Adaptive Multi-Channel Data Dissemination, (3) Adaptive Information Dissemination with Loan-Based Feedback Control, and (4) On Scheduling Sequential Objects with Periodicity for Dynamic Information Dissemination. Considering the nature of dynamic traffic changes, Chapter 2 devises an innovative traffic awareness mechanism, by exploiting the significant potential of client impatience, to estimate the dynamic access frequency distribution. In comparison with the probing and the feedback techniques, the proposed \textit{selective deferment and reflection} (SDR) technique is light-weight and of low complexity. Extensive simulations show that in the case of an increasing/decreasing workload, the real access frequency distribution is bounded by two specific estimated distributions. This fact in turn suggests us to devise a trigonometric tuning method to further enhance the estimation. Consequently, the dynamic traffic awareness mechanism is able to generate an access frequency distribution very close to the real one, showing the feasibility and reliability for adaptive data broadcasting. Chapter 3 proposes an \textit{adaptive multi-channel data dissemination} (AMD) mechanism, which supports the \textit{multi-channel traffic awareness} (MCTA) and the \textit{deterministic balance search} (DBS), in order to pursuit the fairness and robustness for a hybrid data delivery in multi-channel data dissemination environments. In light of the SDR technique, MCTA\ is able to periodically estimate the dynamic access frequencies of all items in the push channels. With the MCTA estimation, furthermore, DBS employs a heuristic search for the global balance, where the push access time and the pull response time are minimized. The experimental results show that the estimated access frequency distribution by MCTA has high accuracy. In addition, DBS\ is robust against the slight accuracy difference caused by MCTA, and notedly attains a global balance very close to the optimum. In order to support the dissemination of static and dynamic information services simultaneously, the work in Chapter 4 devises the \textit{% group-based information dissemination} (GID) mechanism with the \textit{% loan-based slot allocation and feedback control} (LSAFC) technique. The GID mechanism not only supports dynamic information dissemination, but also is backward compatible with the conventional data broadcast scheme. Both of the client-oriented and the server-oriented traffic factors are involved in the GID\ mechanism. Further, we design the LSAFC technique to sustain dynamic traffic changes. Accordingly, the integration of the GID\ mechanism and the LSAFC\ technique is able to perform the adaptations of broadcast slot allocation, group association, group popularity and classification, and broadcast program. Performance study shows that the GID mechanism is very scalable and attains a substantial reduction of dynamic message traffic. In addition, the LSAFC technique is able to functionally complement the GID mechanism, therefore making the GID mechanism more efficient and robust. With regard to data association, dependency, and dynamics, Chapter 5 studies the problem of scheduling dynamic time-sequential data objects in a generalized clients-providers-servers system. According to the principle of periodicity, the scheduling optimization is transformed as the procedure of finding a solution in solving the algebraic problem. In light of the derived upper and lower bounds of mean service access time, a deterministic selection algorithm is developed for the optimal approximation. Several gain measure functions are designed to incrementally enhance the deterministic object selection in response to dynamic data traffic. The broadcast sequence is guaranteed corresponding to the object production order so as to make sure of the correctness of information in the broadcast schedule. Performance evaluation shows the presented deterministic strategy performs prominently in the approximation of schedule optimization. Hence, the mean service access time is minimized to the extent of being close to the theoretical optimum. Ming-Syan Chen 陳銘憲 2003 學位論文 ; thesis 145 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
author2 |
Ming-Syan Chen |
author_facet |
Ming-Syan Chen Chih-Lin Hu 胡誌麟 |
author |
Chih-Lin Hu 胡誌麟 |
spellingShingle |
Chih-Lin Hu 胡誌麟 Dissemination of Dynamic Data and Information Services on Broadcast Channels |
author_sort |
Chih-Lin Hu |
title |
Dissemination of Dynamic Data and Information Services on Broadcast Channels |
title_short |
Dissemination of Dynamic Data and Information Services on Broadcast Channels |
title_full |
Dissemination of Dynamic Data and Information Services on Broadcast Channels |
title_fullStr |
Dissemination of Dynamic Data and Information Services on Broadcast Channels |
title_full_unstemmed |
Dissemination of Dynamic Data and Information Services on Broadcast Channels |
title_sort |
dissemination of dynamic data and information services on broadcast channels |
publishDate |
2003 |
url |
http://ndltd.ncl.edu.tw/handle/34319669630449783038 |
work_keys_str_mv |
AT chihlinhu disseminationofdynamicdataandinformationservicesonbroadcastchannels AT húzhìlín disseminationofdynamicdataandinformationservicesonbroadcastchannels AT chihlinhu guǎngbōpíndàoshàngdòngtàizīliàoyǔzīxùnfúwùchuánbō AT húzhìlín guǎngbōpíndàoshàngdòngtàizīliàoyǔzīxùnfúwùchuánbō |
_version_ |
1718310641124507648 |
description |
博士 === 國立臺灣大學 === 電機工程學研究所 === 91 === The needs of various classes of modern data and information applications,
especially those targeted at mobile computing, demand the support of
adaptive data broadcasting, and the support in turn demands a new kind of
data dissemination framework. For proper implementation, the service model
of adaptive data broadcasting must provide the methodologies of dynamic
traffic awareness and dynamic broadcast content adaptation. At the same
time, these methodologies must be implemented as a distributed data
dissemination platform, having the guarantees of efficiency, reliability,
robustness, and scalability.
We advocate that the design and development of adaptive data broadcasting
schemes should be backward compatible, rather than replacing the
conventional paradigm, in accordance with the principles of information
dissemination and dynamic data management. To this end, in this
dissertation, we formulate the fundamentals of adaptive data broadcasting,
relate them to the traditional data broadcast scheme, and devise several
promising mechanisms. Note that prior researches in the data broadcast realm
are mainly based on the traditional data management model, where data items
are independent, persistent, and static. In fact, however, many modern
information broadcast applications enable the generation and spread of
dynamic data. To fulfil the requirements for the dissemination of dynamic
data and information broadcast services on broadcast channels, we address
the impacts of dynamic traffic awareness and dynamic broadcast content
adaptation. Accordingly, we propose in this dissertation four promising
mechanisms for this purpose:\ particularly, (1) Dynamic Broadcast Traffic
Awareness, (2) Adaptive Multi-Channel Data Dissemination, (3) Adaptive
Information Dissemination with Loan-Based Feedback Control, and (4) On
Scheduling Sequential Objects with Periodicity for Dynamic Information
Dissemination.
Considering the nature of dynamic traffic changes, Chapter 2 devises an
innovative traffic awareness mechanism, by exploiting the significant
potential of client impatience, to estimate the dynamic access frequency
distribution. In comparison with the probing and the feedback techniques,
the proposed \textit{selective deferment and reflection} (SDR) technique is
light-weight and of low complexity. Extensive simulations show that in the
case of an increasing/decreasing workload, the real access frequency
distribution is bounded by two specific estimated distributions. This fact
in turn suggests us to devise a trigonometric tuning method to further
enhance the estimation. Consequently, the dynamic traffic awareness
mechanism is able to generate an access frequency distribution very close to
the real one, showing the feasibility and reliability for adaptive data
broadcasting.
Chapter 3 proposes an \textit{adaptive multi-channel data dissemination}
(AMD) mechanism, which supports the \textit{multi-channel traffic awareness}
(MCTA) and the \textit{deterministic balance search} (DBS), in order to
pursuit the fairness and robustness for a hybrid data delivery in
multi-channel data dissemination environments. In light of the SDR
technique, MCTA\ is able to periodically estimate the dynamic access
frequencies of all items in the push channels. With the MCTA estimation,
furthermore, DBS employs a heuristic search for the global balance, where
the push access time and the pull response time are minimized. The
experimental results show that the estimated access frequency distribution
by MCTA has high accuracy. In addition, DBS\ is robust against the slight
accuracy difference caused by MCTA, and notedly attains a global balance
very close to the optimum.
In order to support the dissemination of static and dynamic information
services simultaneously, the work in Chapter 4 devises the \textit{%
group-based information dissemination} (GID) mechanism with the \textit{%
loan-based slot allocation and feedback control} (LSAFC) technique. The GID mechanism not only supports dynamic information dissemination, but also is
backward compatible with the conventional data broadcast scheme. Both of the
client-oriented and the server-oriented traffic factors are involved in the
GID\ mechanism. Further, we design the LSAFC technique to sustain dynamic
traffic changes. Accordingly, the integration of the GID\ mechanism and the
LSAFC\ technique is able to perform the adaptations of broadcast slot
allocation, group association, group popularity and classification, and
broadcast program. Performance study shows that the GID mechanism is very
scalable and attains a substantial reduction of dynamic message traffic. In
addition, the LSAFC technique is able to functionally complement the GID
mechanism, therefore making the GID mechanism more efficient and robust.
With regard to data association, dependency, and dynamics, Chapter 5 studies
the problem of scheduling dynamic time-sequential data objects in a
generalized clients-providers-servers system. According to the principle of
periodicity, the scheduling optimization is transformed as the procedure of
finding a solution in solving the algebraic problem. In light of the derived
upper and lower bounds of mean service access time, a deterministic
selection algorithm is developed for the optimal approximation. Several gain
measure functions are designed to incrementally enhance the deterministic
object selection in response to dynamic data traffic. The broadcast sequence
is guaranteed corresponding to the object production order so as to make
sure of the correctness of information in the broadcast schedule.
Performance evaluation shows the presented deterministic strategy performs
prominently in the approximation of schedule optimization. Hence, the mean
service access time is minimized to the extent of being close to the
theoretical optimum.
|