Myro Sea™

Do research, enjoy life in meantime

About


PhD Committee
UCSB, ECE Dept
- Prof. Louise Moser (chair),
- Prof. Melliar Smith,
- Prof. Volkan Rodoplu,
UCSB, CS Dept
- Prof. Jianwen Su (苏建文),
Industry
- Jim Kirkley (QAD)



Awards

e-Healthcare
1st place in the 2nd IEEE Services Computing Contest in Salt Lake City, July 2007

MIDAS
3rd place in the 1st IEEE Services Computing Contest in Chicago, Sept 2006

Outstanding TA
Given by Engineering department, during Winter 2006 TAShip for Network Computing (ECE155B) class

I have strong background in advanced Web technologies, distributed systems, databases. These are the topics and conference/journal papers that I was involved during my PhD studies. I am motivated to take responsibilities to lead a large scale project involving these subjects.

2009

A Distributed E-Healthcare System

[Handbook09] Handbook of Research on Distributed Medical Informatics and E-Health, 2009

In this chapter we describe a distributed e-healthcare system that uses the Service Oriented Architecture as a basis for designing, implementing, deploying, invoking and managing healthcare services. The e-healthcare system that we have developed provides support for patients, physicians, nurses, pharmacists and other healthcare professionals, as well as for medical monitoring devices, such as blood pressure monitors. The system transmits e-prescriptions from physicians to pharmacists over the Internet. It offers multi-media input and output, including text, images and speech, to provide a human-friendly interface, with the computers and networks hidden from the user.

Available as google book here

Collaborative Web Data Record Extraction

[ICW09] IEEE International Conference on Web Services

This paper describes a Web Service that automatically parses and extracts data records from Web pages containing structured data. The Web Service allows multiple users to share and manage a Web data record extraction task to increase its utility. A recommendation system, based on the Probabilistic Latency Semantic Indexing algorithm, enables a user to find potentially interesting content or other users who share the same interests with the user. A distributed computing platform improves the scalability of the Web Service in supporting multiple users by employing multiple server computers. A Web Service interface allows users to access the Web Service, and allows programmers to develop their own applications and, thus, extend the functionality of the Web Service.
2008

Building a Distributed E-Healthcare System Using SOA

[ITPro08] IT Professional, Issue on Healthcare, March/April 2008

The healthcare industry in the USA presents special problems for electronic record keeping and communication because of the highly diverse and distributed nature of healthcare. Our Service Oriented Architecture, along with Web Services and Atom / RSS, allows healthcare providers to use diverse computer systems while enabling communication of healthcare records and prescriptions. This article describes a distributed e-healthcare system that uses the Service Oriented Architecture as a means of designing, implementing, deploying, invoking and managing healthcare services. The e-healthcare system provides support for physicians, nurses, pharmacists and other healthcare professionals, as well as for patients and medical devices used to monitor patients. Multi-media input and output, with text, images and speech, make the system more human friendly than if only text is used.

A Collaborative Computing Infrastructure Based on Atom

[CTS08] International Symposium on Collaborative Technologies and Systems

Collaborative computing infrastructure (CCI) facilitates interactions between distributed applications using the pull based, event based model of distributed computing. The Collaborative Computing Infrastructure uses the Atom XML format to provide interoperability between application programs, even if they run on dissimilar computing platforms. The collaborating applications can be programmed in different programming languages and can use different data schema. The Atom Server allows applications to publish their output as Atom feed entries for other applications to read using the Atom Publishing Protocol. We have implemented the currently specified Atom Publishing Protocol, which requires feed entries to be retrieved one by one, and also a modified version of the Atom Publishing Protocol, which supports retrieval of all of the feed entries in a collection at once. The modified version results in shorter retrieval times for feed entries, requires fewer connections to the Atom Server for each consumer, and supports larger numbers of concurrent consumers.

Collaborative Computing Using the Atom Publishing Protocol

[ITNG08] International Conference on Information Technology

Collaborative computing infrastructure (CCI) enables applications that are distributed across a network of computers to collaborate using the Atom Publishing Protocol. The Collaborative Computing Infrastructure couples together, and facilitates interactions between, application programs using the pull-based, event-based distributed computing paradigm. The Atom Server allows publishers to publish events as Atom feed entries for consumers to read. The current specification of the Atom Publishing Protocol, which we have implemented, requires feed entries to be retrieved one by one. To provide better performance, we have also implemented a modification of the Atom Publishing Protocol that supports retrieval of all of the feed entries in a collection at once. This modification results in shorter retrieval times for feed entries, fewer connections to the Atom Server for each consumer, and larger numbers of concurrent consumers.

Presentation

A Reservation-Based Extended Transaction Protocol for Coordination of Web Services

[WSRJ08] International Journal of Web Services Research

Web Services can be used to automate business activities that span multiple enterprises over the Internet. Such business activities can take a significant amount of time to complete, and require a coordination protocol to reach consistent results among the several participants in the business activity. In the current state-of-the-art, either classical distributed transactions or extended transaction protocols with compensation transactions are used. However, classical distributed transactions lock data in the databases of different enterprises for unacceptable durations or involve repeated retries, and compensation transactions can lead to inconsistencies in the databases of the enterprises. Reservation based coordination protocol is a novel extended transaction protocol that can be used to coordinate the tasks of a business activity. Instead of resorting to compensation transactions, our Reservation Protocol employs an explicit reservation phase and an explicit confirmation/cancellation phase. Reservation Protocol maps to the Web Services Coordination specification.

2007

Database Fusion using Atom

[ISCIS07] International Symposium on Computer and Information Sciences

Database fusion produces a local database consisting of aggregated information from multiple remote databases that is selected and transformed into a common table format and data representation. The Database Fusion infrastructure provides Reliable Data Distribution (RDD), Consistent Data Replication (CDR), and Database Aggregation (DBA) and is based on the Atom syndication technology. RDD ensures that the publisher knows that a consumer has received the data intended for it and, thus, that the publisher can garbage collect the data. CDR uses RDD to provide data replication from a source database to target database(s) for the purposes of security, high availability and fast local access. DBA uses CDR to aggregate information from different database sources, possibly deployed on different hosts and on different platforms, and using different database schemas.

Presentation

A Distributed e-Healthcare System

[SCC07] IEEE International Conference on Services Computing

Large-scale distributed systems, such as healthcare systems, are difficult to develop due to their complex and decentralized nature. The Service Oriented Architecture facilitates the development of such systems by supporting modular design, application integration and interoperation, and collaboration between independent software systems. Using open standards, such as XML, SOAP, WSDL and UDDI, the Service Oriented Architecture supports interoperability between services operating on different platforms, and between applications implemented in different programming languages. We designed and implemented distributed e-healthcare system that uses the Service Oriented Architecture as a basis for designing, implementing, deploying, invoking and managing healthcare services. The e-healthcare system that we have developed provides support for physicians, nurses, pharmacists and other healthcare professionals, as well as for patients and medical devices used for monitoring patients. Multi-media input and output, with text, images, speech and keyboard, make the system more user friendly than existing healthcare systems.

Presentation

RDD and CDR

[ICOMP07] International Conference on Internet Computing

Atom is a lightweight syndication technology, based on XML, that allows data to be published on, and retrieved from, the Web. Atom does not currently provide reliable data distribution or consistent data replication. We designed and implemented Reliable Data Distribution (RDD) protocol and Consistent Data Replication (CDR) Protocol. Reliable Data Distribution ensures that the intended consumers have obtained the data that the publisher published and that the publisher can garbage collect the data. Consistent Data Replication provides high availability and fast local access to the data at the consumers.

Presentation

2006

MIDAS: An Automated Supply Chain Management

[SCC06] IEEE International Conference on Services Computing

The MIDAS system that we have developed is an automated supply chain management system based on the Service Oriented Architecture and Web Services. MIDAS operates in a loosely-coupled distributed environment that allows customers, manufacturers, and suppliers to cooperate over the Internet and World Wide Web. It aims to reduce inventory carrying costs and logistics administration costs, yielding a more efficient supply chain, by supporting on-demand, just-in-time manufacturing. MIDAS substantially reduces human intervention on both the customer/manufacturer side and the manufacturer/supplier side. The manufacturer can use one of several strategies to aggregate customers' orders before it processes them and to accumulate suppliers' quotes before it decides on a particular supplier. For these strategies, we evaluate the customer's satisfaction, as measured by the customer response time, and the manufacturer's gain, as measured by the number of orders aggregated or the best price ratio of orders.

Presentation

Publications

Donations