Mobile Computing - Research Survey May 05 2012
Download as PPTX, PDF2 likes388 views
This document summarizes research challenges in mobile computing. It discusses fields like data management, security, data mining, and distributed databases. For each area, it identifies challenges like limited resources, dynamic connectivity, and lack of centralization. It proposes solutions like context-aware and collaborative systems, distributed intrusion detection, and algorithms for trust and authentication in mobile environments. The goal is to develop intelligent systems that can adapt to challenges in mobile and pervasive computing domains.
Mobile Computing - Research Survey May 05 2012
- 1. Mobile Computing: Research Survey UMBC Spring 2012 05/05/2012 Mobile Computing: Research Survey 1 - Joseph Hennawy Research Advisor: Dr. Anupam Joshi
- 2. UMBC Spring 2012 05/05/2012 Mobile Computing: Research Survey 2 Agenda • Introduction • Mobile Computing Fields of Challenge • Data Management • Security • Data Mining • Distributed Databases • Conclusion & Potential Future Research
- 3. UMBC Spring 2012 05/05/2012 Mobile Computing: Research Survey 3 Introduction The Uniqueness of the mobile environment • Constrained Communications Capability • Limited Power Resources • Frequent Disconnects • Asymmetric Communications • Dynamic System Topology • Limited Node Capabilities
- 4. 05/05/2012 UMBC Spring 2012 4Mobile Computing: Research Survey Knowledge Discovery in ubiquitous environments • Ubiquitous Technologies used • Resource Awareness • Ubiquities data collection • Ubiquities data computing
- 5. 05/05/2012 UMBC Spring 2012 5Mobile Computing: Research Survey Knowledge Discovery in ubiquitous environments • Security and Privacy • Human Computer Interface (HCI) • Application Area
- 6. 05/05/2012 UMBC Spring 2012 6Mobile Computing: Research Survey Data Management challenges In distributed computing three main architectures are defined for exchanging data between system elements: • Client-server interaction model • Client-proxy-server interaction model • Peer-to-peer interaction model
- 7. 05/05/2012 UMBC Spring 2012 7Mobile Computing: Research Survey Data Management challenges • Networking Challenges • Device discovery • Routing • RF and wireless links
- 8. 05/05/2012 UMBC Spring 2012 8Mobile Computing: Research Survey Data Management challenges • Context Awareness Challenges • Service and Data Discovery • Mobile location maintenance of network nodes. • User context and profiling
- 9. 05/05/2012 UMBC Spring 2012 9Mobile Computing: Research Survey Data Management challenges • Distributed Systems Data Management Challenges • Distributed Processing and threading • Distributed systems messaging and communication • Naming name spaces, and name resolution • Synchronization • Consistency and Replication • Fault tolerance
- 10. 05/05/2012 UMBC Spring 2012 10Mobile Computing: Research Survey Developing intelligent Mobile Data Management systems • Cross layer intelligent collaboration • Collaborative intelligence • Context aware intelligence
- 11. 05/05/2012 UMBC Spring 2012 11Mobile Computing: Research Survey Security challenges • No prior knowledge of participating nodes due to the dynamic mature of the system. • At any given point, the composition of the system can be diverse enough that a centralized node identification, and authentication is not feasible. • The possibly fragmentary nature of the system also makes it harder to perform security tasks as monitoring, intrusion detection, etc. • Mobile Systems wireless communications technologies are broadcast in nature.
- 12. 05/05/2012 UMBC Spring 2012 12Mobile Computing: Research Survey Security challenges • Wireless communications physical characteristics of such networks make them vulnerable to security threats such as intrusion, denial of service, masquerading, and tampering. • Limited processing, memory, storage and power resources make it harder to implement resource intensive security processes and functions. • The collaborative nature of these systems forces them to disseminate valuable information about system components and data, without central routing control, or monitoring.
- 13. 05/05/2012 UMBC Spring 2012 13Mobile Computing: Research Survey Intrusion Detection in Mobile Systems • Traditionally intrusion detection techniques fall short in such environment due to: • Higher rates of false positives or vice versa due to the unreliable RF communications environment • Lack of consistent tracking mechanisms due to the dynamic nature of the system. • Lack of a solid definition of a misbehaving node, or fixing a reputation associated with a given node, due to the heterogeneous nature of the system, its communications mechanisms and protocols. • Lake of definition or the existence of a single administrative domain. • No global view of the networked system, and the reliance on approximate and localized observations on making intrusion detection mechanisms.
- 14. 05/05/2012 UMBC Spring 2012 14Mobile Computing: Research Survey Intrusion Detection in Mobile Systems • There is a need for a Distributed Intrusion Detection Systems (DIDS) that fits the mobile environment. • Security system is envisioned to use light weight collaborative data mining of various network and distributed system communications stack. • Mining activities will produce, enhance, and match patterns of attacks, intrusions, misbehavior of particulate nodes. • These patterns will be fed into machine learning intelligent agents that will reason, adapt, and enhance, and ask for more mining activities for a system, or a system segment security.
- 15. 05/05/2012 UMBC Spring 2012 15Mobile Computing: Research Survey Node identification, authentication & privileges • The assumption of an existing central authentication, and resource privilege granting entity that does not suffer from communication reachability issues is not valid. • Research performed for devices to compute trust and beliefs about neighboring peers and share them to produce a global trust and belief picture. • Research performed in implementing an authentication framework of where a delegation scheme of authentications, privileges and roles can be done in a mobile environment to accommodate nodes joining a system, or a system segment.
- 16. 05/05/2012 UMBC Spring 2012 16Mobile Computing: Research Survey Node identification, authentication & privileges “ We believe that both trust modeling and authentication distribution activities can be rolled up in a coherent framework where both segments can interact to provide a back and forth credentials, and trust metrics feedback based on the ongoing member nodes dynamics. The intrusion detection activities described above can enhance the proposed framework to offer and intelligent and adaptable security solution based on mobile data mining, autonomous machine learning, and a distributed credentials and authentication engines.”
- 17. 05/05/2012 UMBC Spring 2012 17Mobile Computing: Research Survey Data Mining Challenges • Data mining combines techniques from a variety of field that includes: • Statistics • Machine Learning • Pattern Recognition • Database Systems • Information Retrieval • World-Wide Web • Visualization • Application Specific Domains related to the data being mined
- 18. 05/05/2012 UMBC Spring 2012 18Mobile Computing: Research Survey Data Mining Challenges • Data mining techniques includes: • Summarization • Sampling • Approximation • Clustering • Pattern Recognition • Classification • Outlier Detection
- 19. 05/05/2012 UMBC Spring 2012 19Mobile Computing: Research Survey Data Mining Challenges • Data Mining Fields related to distributed systems: • Stream Data Mining • Moving Objects Mining • Web and Text Data Mining
- 20. 05/05/2012 UMBC Spring 2012 20Mobile Computing: Research Survey Web and Text Data Mining • Web mining can be decomposed into three main styles of mining: • Web Content Mining • Web Structure Mining • Web Usage Mining
- 21. 05/05/2012 UMBC Spring 2012 21Mobile Computing: Research Survey Web and Text Data Mining • For its textual contents, web mining algorithms uses the common textual data mining techniques such as : • Information extraction • Topic tracking • Summarization • Categorization • Clustering • Concept linkage • Multimedia web contents and its integration with web textual contents is still an open field for web mining research activities with lots of potential promises.
- 22. 05/05/2012 UMBC Spring 2012 22Mobile Computing: Research Survey Distributed Databases Challenge The breakdown of mobile querying
- 23. 05/05/2012 UMBC Spring 2012 23Mobile Computing: Research Survey Distributed Databases Challenge • Mobile databases performance measures in term of reliability, availability, correctness, and timeliness. • The usage of heterogynous mobile communication infrastructures and combining it with using different query strategies. A layer of intelligence that detects available and changing infrastructure characteristics and adapt query strategies. • Smart resource and data discovery, and fuses that with user profiles and policies to produced smart automated or semi- automated queries.
- 24. 05/05/2012 UMBC Spring 2012 24Mobile Computing: Research Survey Distributed Databases Challenge • Sensor network and Mobile databases integration. Such an integrate can offer a local, or global sensor view based on database queries or processes or semi-processed sensor data. • The definition of meta-data needed for multimedia database. Researching the way the methods of meta-data are being produced, the ways that meta-data can model multimedia databases, and expanding the usage of meta-data to be a source for upper level intelligence functions that builds semantics based on them is needed.
- 25. 05/05/2012 UMBC Spring 2012 25Mobile Computing: Research Survey Distributed Databases Challenge • Sensor network and Mobile databases integration. Such an integrate can offer a local, or global sensor view based on database queries or processes or semi-processed sensor data. • The definition of meta-data needed for multimedia database. Researching the way the methods of meta-data are being produced, the ways that meta-data can model multimedia databases, and expanding the usage of meta-data to be a source for upper level intelligence functions that builds semantics based on them is needed.
- 26. 05/05/2012 UMBC Spring 2012 26Mobile Computing: Research Survey