Alexandros Biliris

CDN brokering

Content distribution networks (CDNs) increase the capacity of individual Web sites and attempt to deliver content from caches that are located closer to end-users than the origin servers that provide the content. CDN brokering provides CDNs a way to interoperate by allowing one CDN to intelligently redirect clients dynamically to other CDNs. This paper describes the goals, architecture, and performance of a CDN brokerage system. Our system has been deployed on the Internet on a provisional basis, and our architectural ideas have helped advance the evolution of Internet standards for interoperating CDNs.

Synchronization and recovery in a client-server storage system

Abstract. Client-server object-oriented database management systems differ significantly from traditional centralized systems in terms of their architecture and the applications they target. In this paper, we present the client-server architecture of the EOS storage manager and we describe the concurrency control and recovery mechanisms it employs. EOS offers a semi-optimistic locking scheme based on the multi-granularity two-version two-phase locking protocol. Under this scheme, multiple concurrent readers are allowed to access a data item while it is being updated by a single writer. Recovery is based on write-ahead redo-only logging. Log records are generated at the clients and they are shipped to the server during normal execution and at transaction commit. Transaction rollback is fast because there are no updates that have to be undone, and recovery from system crashes requires only one scan of the log for installing the changes made by transactions that committed before the crash. We also present a preliminary performance evaluation of the implementation of the above mechanisms.n

prospector: a content-based multimedia server for massively parallel architectures

The Prospector Multimedia Object Manager prototype is a general-purpose content analysis multimedia server designed for massively parallel processor environments. Prospector defines and manipulates user defined functions which are invoked in parallel to analyze/manipulate the contents of multimedia objects. Several computationally intensive applications of this technology based on large persistent datasets include: fingerprint matching, signature verification, face recognition, and speech recognition/translation [OIS96].

BeSS object storage manager: architecture overview

BeSS is a high performance, memory-mapped object storage manager offering distributed transaction management facilities and extensible support for persistence. In this paper, we present an overview of the peer-to-peer architecture of BeSS, and we discuss issues related to space management, inter-object references, database corruption, operation modes, cache replacement, and transaction management.

BeSS: storage support for interactive visualization systems

Interactive computer graphics systems for visualization of realistic-looking, three-dimensional models are useful for evaluation, design and training in virtual environments, such as those found in architectural and mechanical CAD, ight simulation, and virtual reality. Interactive visualization systems display images of a threedimensional model on the screen of a computer workstation as seen from a simulated observers viewpoint under control by a user. If images are rendered smoothly and quickly enough, an illusion of real-time exploration of a virtual environment can be achieved as the simulated observer moves through the model. Unfortunately, models of interesting virtual environments (e.g., buildings with furniture) comprise millions of polygons and contain more data (gigabytes) than can t into memory on most graphics workstations. In order to support real-time walkthroughs of such virtual environments, a visualization system must store only a subset of the model in memory at once (i.e., the working set) and swap di erent parts in and out of memory in real-time as the observer navigates through the environment interactively. In order to avoid delays to the perceptible frame rate due to page faults, an interactive walkthrough system must prefetch data asynchronously so that the data is resident in memory before it is displayed.