Carlos Abalde

ARMISTICE: an experience developing management software with Erlang

In this paper, some experiences of using the concurrent functional language Erlang to implement a classical vertical application, a risk management information system, are presented. Due to the complex nature of the business logic and the interactions involved in the client/server architecture deployed, traditional development techniques are unsatisfactory. First, the nature of the problem suggests an iterative design approach. The use of abstractions (functional patterns) and compositionality (both functional and concurrent composition) have been key factors to reduce the amount of time spent adapting the system to changes in requirements. Despite our initial concerns, the gap between classical software engineering and the functional programming paradigm has been successfully fullfiled.

On modelling agent systems with erlang

Multi-agent systems are a kind of concurrent distributed systems. In this work, some guidelines on how to create multi-agent systems using Erlang are presented. The modelled system supports cooperation among agents by plan exchange, reconfiguration and has a certain fault-tolerance. The distributed and concurrent functional programming Erlang, together with OTP platform, allows the creation of high-availability and fault-tolerant concurrent and distributed systems, and it seems to be an interesting framework for implementing multi-agent systems.

Efficient Discovery of Widely Distributed and Non-volatile Resources on DHTs

In recent years, a new generation of structured and decentralized P2P CDNs based on distributed hash tables (DHTs) has emerged. Nowadays, DHT algorithms are a powerful building block when designing distributed behaviors. However, sometimes DHTs are not flexible enough and problems arise when an efficient lookup in the DHT structure itself -instead of in the stored data- is required. Our motivation scenario is the resource discovery problem in a decentralized and distributed caching architecture, deployed in a set of clusters built from cheap off-the-shelf computers organized in a DHT overlay network. In this paper we present the design of a resource discovery service layered on the underlying DHT and based on the combination of(1) an spanning-tree built mapping DHT nodes to their parents in a bottom-up fashion, and (2) a set of routing indices which allow nodes to efficiently lookup other nodes in the DHT overlay network, matching some resource constraints This service does not alter the underlying DHT behavior, it scales to large wide area systems and it tracks both relatively static and frequently changing resources. Furthermore, the results of the experiments conducted to evaluate the good performance and scalability of the architecture are also presented.

Managing the Risks of Risk Management

Enterprise Risk Management (ERM) is a field of enormous importance due to its economic repercussion. More and more companies are paying attention to it, given there is significant economic savings for attending to the various aspects of risk. At the same time, ERM has witnessed a shift in the way firms manage the many uncertainties that stand in the way of achieving their strategic, operational and financial objectives. Nevertheless, existing Risk Management Systems (RMIS) are neither powerful, nor flexible enough to model the complexity of the risk management process. In this paper we will examine these issues in greater detail, explaining the existing problem and proposing a software system that addresses them. We will also show the actual design and operation of ARMISTICE (Advanced Risk Management Information System: Tracking Insurances, Claims and Exposures), a successful RMIS case study.

Lambda Goes to Hollywood

In this paper, some experiences of using the concurrent functional language Erlang to implement a distributed video-on-demand server are presented. This server is the result of a project supported by a regional cable company, and it is intended to provide services for real users in the real world. The nature of the problem, with fuzzy and changing requirements, suggests a highly flexible and scalable architecture. The use of abstractions (functional patterns) and compositionality (both functional and concurrent composition) have been key factors to reduce the amount of time spent adapting the system to changes in requirements. Despite our initial concerns, efficiency constraints have been succesfully met.

A Distributed System for Massive Generation of Synthetic Video Using GPUs

Audio-visual interactive content is very common nowadays. It has many applications in very different fields, from videogames to visualization of scientific data. However, there are environments such as digital television in which the delivery of interactive content is of interest, but are limited by the shortcomings of the players. For example, in cable TV environments users access content through a set-top box, which is usually very limited in computing power due to cost, power consumption and the need to keep a moderate size. Furthermore, set-top boxes do not usually have specific hardware for graphics processing (GPU, Graphics Processing Unit) desirable for high quality interactive content, but rather are optimized for real time decoding of video in hardware (usually Mpeg-2, in very recent ones h.264). In this work we describe a distributed system for the creation of synthetic content and its encoding to digital video to send it to the clients. The most important features to provide are scalability and fault tolerance, in order to support a large number of concurrent users with an uninterrupted service.

Extending the VoDKA Architecture with P2P Aggregated Content Management

This paper presents our results building a decentralized and structured P2P video content distribution network, developed as an extension of an existing on-demand streaming server. It has been designed to improve streaming server capacity, without any infrastructure upgrade, when it must deal with big LAN-networks and highly-correlated server access patterns. A coordination algorithm build on top of a distributed hash table is presented, and its integration inside the streaming server architecture is explained. Experimental results show that the design is an appropriate approach for video content distribution with interesting scalability and availability properties.

Towards a decentralized and structured network of P2P public information screens

This paper presents our results designing a peer-to-peer (P2P) video content distribution network (CDN). Based on a distributed hash table (DHT) algorithm and developed as an extension of an existing on-demand streaming server, this CDN is designed to improve server streaming capacitywithout any infrastructure upgrade to dealwith large networks of information screens. Experimental results show that the design is an appropriate approach for video content distribution with interesting scalable and availability features.

Building functional patterns for implementing distributed applications

Combining the use of design patterns and distributed functional programming is pointed out as a key factor to produce correct distributed systems running on a cluster of computers. A high degree of adaptability, fault tolerance and scalability are obtained at low development costs. Patterns of recurrent distributed design problems are presented and implemented using the distributed functional language Erlang. The design patterns discussed involve coordination mechanisms of concurrent threads such as guarded suspension or scheduling abstractions. The paper includes some applications of the presented patterns and a small experiment with a clustered XSLT engine.

Resource Discovery Approach to Support a QoS-aware DHT-Based Caching Architecture

Though widely accepted as a key building block for next-generation large scale decentralized systems, the lack of flexibility of DHTs on efficient non identifier-based lookups is a well-known problem. In this paper, a resource discovery service that tackles the issues identified in a decentralized and distributed DHT-based caching architecture for media content distribution is presented. In the proposed approach, complex queries on dynamic peer resources can be performed. This discovery service is layered over an underlying DHT overlay network which connects all the peers, based on the combination of a spanningtree built bottom-up by mapping DHT peers to their parents, and routing indices which allow peers to efficiently lookup other peers, matching some resource constraints. Measures from simulation as well as from a real implementation are analyzed.