Éva Hosszu

Physical impairments of monitoring trails in all optical transparent networks

In recent years, following the deployment of WDM networks, fault detection and localization has become a challenging issue in networks with high reliability. Optical layer monitoring schemes based on monitoring trail (m-trail) is considered a promising approach localizing single link failure unambiguously in all optical networks. Although the extensive works on the m-trail concept, the issue has not been validated from the feasibility point of view. Previous works on the m-trail monitoring scheme have focused on algorithm design for minimizing the number of monitors, however, none of them has observed the physical limitations. In this paper, we investigate the physical constraints on launching m-trails, mainly focusing on the maximum length each m-trail may have. Numerous simulations were implemented in VPI Transmission Maker Simulation Tool for observing qualitative parameters in different m-trail length. In our simulation, we could validate the feasibility of 15000 km long m-trails when out-of-band monitoring is used.

A resource-aware and time-critical IoT framework

Internet of Things (IoT) systems produce great amount of data, but usually have insufficient resources to process them in the edge. Several time-critical IoT scenarios have emerged and created a challenge of supporting low latency applications. At the same time cloud computing became a success in delivering computing as a service at affordable price with great scalability and high reliability. We propose an intelligent resource allocation system that optimally selects the important IoT data streams to transfer to the cloud for processing. The optimization runs on utility functions computed by predictor algorithms that forecast future events with some probabilistic confidence based on a dynamically recalculated data model. We investigate ways of reducing specifically the upload bandwidth of IoT video streams and propose techniques to compute the corresponding utility functions. We built a prototype for a smart squash court and simulated multiple courts to measure the efficiency of dynamic allocation of network and cloud resources for event detection during squash games. By continuously adapting to the observed system state and maximizing the expected quality of detection within the resource constraints our system can save up to 70% of the resources compared to the naive solution.

Combinatorial error detection in linear encoders

Linear error correction is widely implemented in millions of telecommunication devices to cope with unreliable or noisy communication channels. A common error in linear encoders is when some bits that are originally one in the generator matrix get erased and become zero - usually due to a soft error becoming a hard error or simply a physical impact on the device. The need to deal with this phenomenon has been recognized in the literature, although to the best of our knowledge no one has ever tried to diagnose erasure-type faults in an encoder. In this paper we focus on the situation when the linear encoder hardware unit becomes faulty and propose a novel combinatorial fault detection scheme based on group testing to diagnose its state with great efficiency in a quick manner. Furthermore we touch upon some solutions to compensate for at both the sender and the receiver side.

On a parity based group testing algorithm

In traditional Combinatorial Group Testing the problem is to identify up to d defective items from a set of n items on the basis of group tests. In this paper we describe a variant of the group testing problem above, which we call parity group testing. The problem is to identify up to d defective items from a set of n items as in the classical group test problem. The main difference is that we check the parity of the defective items in a subset. The test can be applied to an arbitrary subset of the n items with two possible outcomes. The test is positive if the number of defective items in the subset is odd, otherwise it is negative. In this paper we extend Hirschberg et al.’s method to the parity group testing scenario.

Constructions for unambiguous node failure localization in grid topologies

Precise, fast and scalable fault localization is a highly desired feature in all optical mesh networks. The monitoring trail (m-trail) framework has been long in use for centralized failure localization, its capabilities were recently utilized to accommodate the distributed scenario as well. Most of the prior art focused solely on link failures; in this study we step further and analyze node failures only. Node failures are fundamentally different from link failures, and thus a new theoretical framework is developed. In particular, we are interested in the scalability of localizing node failures. From an information theoretic view, with b m-trails at most 2b -1 single failures can be identified. This intuitively leads to a very attractive property that the number of monitors might be logarithmic to the size of the network. Does this logarithmic behaviour holds for real life topologies too? As a step towards the answer, we show tight constructions for both centralized and distributed node failure localization.

Fast failure localization in all-optical networks with length-constrained monitoring trails

Monitoring trails (m-trails) have been extensively studied as an alternative to the conventional link-based monitoring approach by using multi-hop supervisory lightpaths in all-optical networks. However, none of the previous studies have investigated the effect of length constraints upon the m-trail formation, which nonetheless correspond to the failure localization time. This paper addresses the above issue and formulates a new m-trail allocation problem, where the relationship between the number of m-trails versus the maximum hop count is explored. First, the paper investigates the theoretical bounds of allocating m-trails with at most k hops via an optimal group testing construction. Secondly, a novel meta-heuristic approach based on bacterial evolutionary algorithm for solving the length-constrained m-trail allocation problem is introduced. Through extensive simulations the performance gap of the proposed algorithm to the lower bound is presented on a wide diversity of topologies.