TranSquid: Transcoding and Caching Proxy for Heterogenous ECommerce Environments

1
TranSquid Transcoding and Caching Proxy for
Heterogenous E-Commerce Environments

• Maheshwari, A. Sharma, A. Ramamritham, K.
  Shenoy, P. Research Issues in Data Engineering
  Engineering E-Commerce/E-Business Systems, 2002.
  RIDE-2EC 2002. Proceedings. Twelfth International
  Workshop on , 2002 Page(s) 50 -59

Yu-Lun, Liao
2
Outline

• Introduction
• Caching Technology
• TransSquid
• Architecture
• Modules
• Transcoding Results
• Conclusion

3
Introduction

• With the advent of the Wireless Internet
• The growth of mobile device
• Current mobile subscriptions 170 million,
  exceed 1 billion by 2003
• More than 1.5 billion PDAs handsets with
  wireless capabilities at the end of 2004
• The client space becomes heterogeneous
• Many Problems
• Lower bandwidth?higher error rates?frequent
  disconnections?smaller displays

4
Introduction (cont)

• Heterogeneous client
• laptops, desktops, kiosks, automobile browsers,
  cellular phones, pagers, pocket PCs, Palm OS
  devices and other handheld devices
• Each device brings its own requirements
• Caching has increased user satisfaction
  tremendously both in homogeneous heterogeneous
  client environment

5
Caching technology

• Cache technology provides three primary
  characteristics
• Scalability?availability?responsiveness
• For ISPs Content Providers
• Reduces traffic?offer high performance at low
  cost
• All caching algorithms try to maximize the
  probability of a HIT given the limited storage
  space with the cache

6
Heterogeneous Client Environment

• Identification of the client type
• Intermediary proxy end server need to
  understand the CC/PP
• A caching proxy can be designed to
• Recognize multiple variants of a resource and
  categorize them in the cache accordingly
• Understand the CC/PP to facilitate client
  recognition
• Manipulate fidelity or modality (or transcode) of
  variants whenever possible to provide better
  service

7
TransSquid

• A multi-level caching solution
• As in Table 1, there are an ever-increasing
  myriad of device
• Provide for caches for each type of device?
• Lead to low and perhaps ineffective HIT rates

8
TransSquid (cont)

• High Capability Client
• PC?Work Station?Laptop
• Large storage (more than 64MB RAM and 10 GB disk
  space)
• Large screen size (640x480 or above)?media
  support and good processing power
• Medium Capability Client
• PDA?WebTop
• Smaller screen size (320x200)
• Limited colors?less processing power and with
  slow wireless link
• Limited Capability Client
• Mobile phones
• Very low resource
• A Cache HIT would outweigh the less optimal
  solution in tanscoding

9
Partial HIT

• HIT, MISS and the Partial HIT
• The cache contains a higher fidelity variant but
  no object in the cache that the client maps to
  exactly
• The cache sends the object to the transcoding
  module

10
Architecture

• A modular framework
• Multi-level Caching Module
• Transcoding Module
• Client Side Module
• Policy Engine
• Replacement Policy
• FIFO

11
Modules

• Multi-level Caching Module
• A hierarchical model
• Transcoding Module
• Image
• Reducing the dimensions, decreasing the no. of
  color the quality factor of JPEG images
• Performed through the policy engine or by using
  end server directed information
• Text
• Removing unwanted parts like adv. and buttons

Cache level 1
High Capability Client
Cache level 2
Cache level 3
Medium Capability Client
Limited Capability Client
12
Modules (cont)

• Policy Engine
• Provide information for Transcoding Module
• Client Side Module
• The interface between the client and the proxy
• Intercepting the client request and mapping it to
  the correct cache
• Maintains a client specific state

13
Flow of Typical Request

• A client connection is accepted by the client
  side and then the HTTP request is parsed
• The access control are checked
• The client side looks for the requested object in
  the cache
• A cache HIT
• A Partial HIT
• Transcoding according to server or policy engine
• A cache MISS

14
Transcoding Results

• Size minify?subsample
• Fidelity JPEG compress?GIF compress?reduce
  resolution
• Color Content reduce colors?convert to
  grey?convert to b/w

15
Transcoding Results (cont)
16
Conclusion

• end-to-end vs. intermediary approach
• More optimized or less responsive
• TransSquid support both end-to-end intermediary
  
• Provide caching transcoding
• A relative issue replacement in transcoding
  proxy