Jussi Petri Myllymaki

An evaluation of binary xml encoding optimizations for fast stream based xml processing

This paper provides an objective evaluation of the performance impacts of binary XML encodings, using a fast stream-based XQuery processor as our representative application. Instead of proposing one binary format and comparing it against standard XML parsers, we investigate the individual effects of several binary encoding techniques that are shared by many proposals. Our goal is to provide a deeper understanding of the performance impacts of binary XML encodings in order to clarify the ongoing and often contentious debate over their merits, particularly in the domain of high performance XML stream processing.

Implementing a scalable XML publish/subscribe system using relational database systems

An XML publish/subscribe system needs to match many XPath queries (subscriptions) over published XML documents. The performance and scalability of the matching algorithm is essential for the system when the number of XPath subscriptions is large. Earlier solutions to this problem usually built large finite state automata for all the XPath subscriptions in memory. The scalability of this approach is limited by the amount of available physical memory. In this paper, we propose an implementation that uses a relational database as the matching engine. The heavy lifting part of evaluating a large number of subscriptions is done inside a relational database using indices and joins. We described several different implementation strategies and presented a performance evaluation. The system shows very good performance and scalability in our experiments, handling millions of subscriptions with moderate amount of physical memory.

Buddy tracking-efficient proximity detection among mobile friends

Global positioning systems (GPS) and mobile phone networks are making it possible to track individual users with an increasing accuracy. It is natural to ask whether one can use this information to maintain social networks. Here each user wishes to be informed whenever one of a list of other users, called the users friends, appears in the users vicinity. In contrast to more traditional positioning based algorithms, the computation here depends not only on the users own position on a static map, but also on the dynamic position of the users friends. Hence it requires both communication and computation resources. The computation can be carried out either between the individual users in a peer-to-peer fashion or by centralized servers where computation and data can be collected at one central location. In the peer-to-peer model, a novel algorithm for minimizing the number of location update messages between pairs of friends is presented. We also present an efficient algorithm for the centralized model, based on region hierarchy and quadtrees. The paper provides an analysis of the two algorithms, compares them with a naive approach, and evaluates them using the IBM city simulator system.

A function-based access control model for XML databases

XML documents are frequently used in applications such as business transactions and medical records involving sensitive information. Typically, parts of documents should be visible to users depending on their roles. For instance, an insurance agent may see the billing information part of a medical document but not the details of the patients medical history. Access control on the basis of data location or value in an XML document is therefore essential. In practice, the number of access control rules is on the order of millions, which is a product of the number of document types (in 1000s) and the number of user roles (in 100s). Therefore, the solution requires high scalability and performance. Current approaches to access control over XML documents have suffered from scalability problems because they tend to work on individual documents. In this paper, we propose a novel approach to XML access control through rule functions that are managed separately from the documents. A rule function is an executable code fragment that encapsulates the access rules (paths and predicates), and is shared by all documents of the same document type. At runtime, the rule functions corresponding to the access request are executed to determine the accessibility of document fragments. Using synthetic and real data, we show the scalability of the scheme by comparing the accessibility evaluation cost of two rule function models. We show that the rule functions generated on user basis is more efficient for XML databases.

Toward automated large-scale information integration and discovery

The high cost of data consolidation is the key market inhibitor to the adoption of traditional information integration and data warehousing solutions. In this paper, we outline a next-generation integrated database management system that takes traditional information integration, content management, and data warehouse techniques to the next level: the system will be able to integrate a very large number of information sources and automatically construct a global business view in terms of “Universal Business Objects”. We describe techniques for discovering, unifying, and aggregating data from a large number of disparate data sources. Enabling technologies for our solution are XML, web services, caching, messaging, and portals for real-time dashboarding and reporting.

Buddy tracking - efficient proximity detection among mobile friends

Global positioning systems (GPS) and mobile phone networks make it possible to track individual users with an increasing accuracy. It is natural to ask whether this information can be used to maintain social networks. In such a network each user wishes to be informed whenever one of a list of other users, called the users friends, appears in the users vicinity. In contrast to more traditional positioning based algorithms, the computation here depends not only on the users own position on a static map, but also on the dynamic position of the users friends. Hence it requires both communication and computation resources. The computation can be carried out either between the individual users in a peer-to-peer fashion or by centralized servers where computation and data can be collected at one central location. In the peer-to-peer model, a novel algorithm for minimizing the number of location update messages between pairs of friends is presented. We also present an efficient algorithm for the centralized model, based on region hierarchy and quadtrees. The paper provides an analysis of the two algorithms, compares them with a naive approach, and evaluates them on user motions generated by the IBM City Simulator system.

High-performance spatial indexing for location-based services

Much attention has been accorded to Location-Based Services and location tracking, a necessary component in active, trigger-based LBS applications. Tracking the location of a large population of moving objects requires very high update and query performance of the underlying spatial index. In this paper we investigate the performance and scalability of three main-memory based spatial indexing methods under dynamic update and query loads: an R-tree, a ZB-tree, and an array/hashtable method. By leveraging the LOCUS performance evaluation testbed and the City Simulator dynamic spatial data generator, we are able to demonstrate the scalability of these methods and determine the maximum population size supported by each method, a useful parameter for capacity planning by wireless carriers.

DynaMark: A Benchmark for Dynamic Spatial Indexing

We propose a performance benchmark for dynamic spatial indexing that is directly geared towards Location-Based Services (LBS). A set of standard, realistic location trace files is used to measure the update and query performance of a spatial data management system. We define several query types relevant for LBS: proximity queries (range queries), k-nearest neighbor queries, and sorted-distance queries. Performance metrics are defined to quantify the cost (elapsed time) of location updates, spatial queries, and spatial index creation and maintenance.

Web Data Extraction Techniques and Applications Using the Extensible Markup Language (XML)

The driving force behind the technology revolution has always been just one thing: information. Almost every invention related to the computer since the transistor has been made to aid in the transferring of a piece of information, or data, from one place to another. Despite the existence of a primitive form of what we now know of as the Internet, less than one generation ago digital information mostly needed to be carried around on magnetic devices such as tapes and disks. Fortunately, the prominent rise of the Internet and the World Wide Web in the mid-1990s removed the barrier that physical transportation of data placed on us.