Distributed Computing

Introduction for second edition /  Preface

When in 1989 an anonymous reviewer commented on my short paper that “this classification should be extended to description of distributed systems,” (Yet another approach to classification of redundancy, CIM IMEKO Symp.1990, Helsinki, pp. 117-124) I was really excited, because people in the research community were thinking much deeper and wider than myself (- I had just defended my PhD).

Further, fault tolerance was migrating to dependability (Jean Claude Laprie was an indisputable authority and expert in this domain, see more:  www.springer.com/gb/book/9783709191729, which later emerged as the concept of resilience.

In principle, all these new properties had a concrete reasoning and meaning behind them: when something erroneous has happens, any system of our design should be able to cope with the problem. Options vary, as well as circumstances and area of application, thus:

- If it stops the error propagating and freezes in a safe state, it is fail-stop, or fail-safe;
- If it can cope with permanent faults inside the system, it is a fault tolerant system;
- When it continues with reduced functionality, it is graceful degradation;
- If it is designed with attention having been paid to reliability, availability and maintenance or serviceability, it is
      dependable system;
- If it capable to of tolerating obstacles caused by internal and external factors and can spring back, recover and continue, then a system can be considered as resilient.

There are two major ways to achieve any of the properties mentioned above: at  system level, or at local level (technological). Obviously, any reasonable combination of both levels is also welcome. In this book we do not want to repeat our papers and books
(https://www.springer.com/gb/book/9783319150680, https://www.springer.com/gb/book/9783319468129)
but to incorporate into the second edition any significant progress that has emerged.
Speaking about ICT systems, especially safety-critical and real-time ones, we might think about the implementation of resilience from the system level down through to hardware and systems software. In addition, we need to consider that each of the parts will both interact with and support each other.

NFRs (Non-Functional Requirements) of each part of the system were considered, such as:

- Performance;
- Reliability;
- Efficiency (mostly energy efficient).

Therefore, the systems that we design should be PRE-smart and provide these properties throughout the life cycle. 

Neither of  our books to date  - have appeared as complete. These books have been used in China, Switzerland, Russia and USA, (mostly Masters and PhD students) and we have received substantial feedback, such as:

- While reliability of hardware and availability at the system level are explained and fine, there are no sections, or chapters about performance, especially where parallel and distributed systems are concerned;
- How to apply (as mentioned in the above review) the classification and properties of resilience for and within distributed systems;
- How real-time and safety-critical applications should be treated considering the system resilience: rules for system and for packages - have they changed?

It was especially satisfying as  we have  discovered that these segments are being updated by researchers around the globe, providing excellent contributions to the content.

Thus, our book became and evolving system in itself, aggregating our further efforts with the efforts and results of our colleagues from China, Switzerland, UK and Russia. Our book has therefore become itself resilient, benefiting from the contributions from the following:

Performance chapter (including element level performance and parallel design was prepared and included using materials and having contributions from:

-  Professor Hao Kai, Shantou University, China;
- Simon Monkman, IT-ACS Ltd researcher.

System software chapters were part of substantial efforts from:

- Professor Eugeny Zuev and his team in Technopolis, Kazan, Russia.

In turn, requested in 1989 consideration of system level of resilience for distributed systems were developed as two chapters: system level and algorithmic implementation prepared by me and Stephen Farrell. In these chapters we have introduced a concept of desperation ( for transactions within distributed systems) and show that our existing and new results, even patented: https://www.ipo.gov.uk/p-ipsum/Case/PublicationNumber/GB2448351

can be extremely useful making the whole network really resilient and achieving by far better service for applications, especially when critical level of their use was assumed.

I Schagaev

Basic similarities among the several service-registry standards that exist today suggest the potential for interoperability. Given that service registries are key components inWeb services architectures, interoperability will play an increasingly important role in facilitating dynamic ...

Cloud computing offers an innovative business model to enterprise for IT services consumption and delivery. Software as a Service (SaaS) is one of the cloud offerings that attract organisations as a potential solution in reducing their IT cost. However, the vast diversity among the available cloud SaaS services makes it difficult for customers to decide whose vendor services to use or even to determine a valid basis for their selections. Moreover, this variety of cloud SaaS services has led to proprietary architectures and technologies being used by cloud vendors, increasing the risk of vendor lock-in for customers. Therefore, when enterprises interact with SaaS providers within the purview of the current cloud marketplace, they often encounter significant lock-in challenges to migrating and interconnecting cloud. Hence, the complexity and variety of cloud SaaS service offerings makes it imperative for businesses to use a clear and well understood decision process to procure, migrate and/or discontinue cloud services. To date, the expertise and technological solutions to simplify such transition and facilitate good decision making to avoid lock-in risks in the cloud are limited. Besides, little investigation has been carried out to provide a comprehensive decision framework to support enterprises on how to avoid lock-in risks when selecting and implementing cloud-based SaaS solutions within existing environments. Such decision framework is important to reduce complexity and variations in implementation patterns on the cloud provider side, while at the same time minimising potential switching cost for enterprises by resolving integration issues with existing IT infrastructures. This paper proposes a holistic 6-step decision framework that enables an enterprise to assess its current IT landscape for potential SaaS replacement, and provides effective strategies to mitigate vendor lock-in risks in cloud (SaaS) migration. The framework follows research findings and addresses the core requirements for choosing vendor-neutral interoperable and portable cloud services without the fear of vendor lock-in, and architectural decisions for secure SaaS migration. Therefore, the results of this research can help IT managers have a safe and effective migration to cloud computing SaaS environment.

Blockchain is a novel technology that is rising a lot of interest in the industrial and re- search sectors because its properties of decentralisation, immutability and data integrity. Initially, the underlying consensus mechanism has been designed for permissionless block- chain on trustless network model through the proof-of-work, i.e. a mathematical challenge which requires high computational power. This solution suffers of poor performances, hence alternative consensus algorithms as the proof-of-stake have been proposed.
Conversely, for permissioned blockchain, where participants are known and authenti- cated, variants of distributed consensus algorithms have been employed. However, most of them comes out without formal expression of security analysis and trust assumptions because the absence of an established knowledge. Therefore the lack of adequate analysis on these algorithms hinders any cautious evaluation of their effectiveness in a real-world setting where systems are deployed over trustless networks, i.e. Internet.
In this thesis we analyse security and performances of permissioned blockchain. Thus we design a general model for such a scenario in a way to propose a general benchmark for the experimental evaluations. This work brings two main contributions.
The first contribution concern the analysis of Proof-of-Authority, a Byzantine Fault- Tolerant consensus protocol. We compare two of the main algorithms, named Aura and Clique, with respect the well-established Practical Byzantine Fault-Tolerant, in terms of se- curity and performances. We refer the CAP theorem for the consistency, availability and partition tolerance guarantees and we describe a possible attack scenario in which one of the algorithms loses consistency. The analysis advocates that Proof-of-Authority for permissioned blockchains deployed over WANs experimenting Byzantine nodes, do not provide adequate consistency guarantees for scenarios where data integrity is essential. We claim that actually the Practical Byzantine Fault-Tolerant can fit better for permis- sioned blockchain, despite a limited loss in terms of performance.
The second contribution is the realisation of a benchmark for practical evaluations. We design a general model for permissioned blockchain under which benchmarking perfor- mances and security guarantees. However, because no experiment can verify all the pos- sible security issues permitted by the model, we prototype an adversarial model which simulate three attacks, feasible for a blockchain system. We then integrate this attacker model in a real blockchain client to evaluate the resiliency of the system and how much the attacks impact performances and security guarantees.

We present the development and use of a novel distributed geohazard modeling environment for the analysis and interpretation of large scale earthquake data sets. Our work demonstrates, for the first time, how earthquake-related surface deformation measured from satellite images using imageodesy algorithms is coupled with analysis and simulation using finite-element numerical models.

Task intensive electronic control units (ECUs) in automotive domain, equipped with multicore processors , real time operating systems (RTOSs) and various application software, should perform efficiently and time deterministically. The parallel computational capability offered by this multicore hardware can only be exploited and utilized if the ECU application software is parallelized. Having provided with such parallelized software, the real time operating system scheduler component should schedule the time critical tasks so that, all the computational cores are utilized to a greater extent and the safety critical deadlines are met. As original equipment manufacturers (OEMs) are always motivated towards adding more sophisticated features to the existing ECUs, a large number of task sets can be effectively scheduled for execution within the bounded time limits. In this paper, a hybrid scheduling algorithm has been proposed, that meticulously calculates the running slack of every task and estimates the probability of meeting deadline either being in the same partitioned queue or by migrating to another. This algorithm was run and tested using a scheduling simulator with different real time task models of periodic tasks. This algorithm was also compared with the existing static priority scheduler, which is suggested by Automotive Open Systems Architecture (AUTOSAR). The performance parameters considered here are, the % of core utilization, average response time and task deadline missing rate. It has been verified that, this proposed algorithm has considerable improvements over the existing partitioned static priority scheduler based on each performance parameter mentioned above.

Emerging pervasive information and computational environments require a content-based middleware infrastructure that is scalable, self-managing, and asynchronous. In this paper, we propose associative rendezvous (AR) as a paradigm for content-based decoupled interactions for pervasive ...

This paper provides a complete framework for thread migration using JPDA. In our framework neither we lose portability nor do we insert any artificial code. The system developed based on our framework needs no extra involvement of programming to continue the whole migration system. It is a powerful autonomous system for heterogeneous environment maintaining its portability since Java Virtual Machine (JVM) is not modified.

We propose a steganalytic algorithm for triangle meshes, based on the supervised training of a classifier by discriminative feature vectors. After a normalization step, the triangle mesh is calibrated by one step of Laplacian smoothing and then a feature vector is computed, encoding geometric information corresponding to vertices, edges and faces. For a given steganographic or watermarking algorithm, we create a training set containing unmarked meshes and meshes marked by that algorithm, and train a classifier using Quadratic Discriminant Analysis. The performance of the proposed method was evaluated on six well-known watermarking/steganographic schemes with satisfactory accuracy rates.

Even though listening comprehension is an important language skill in second language learning, it is considered to be the most difficult skill to learn. Besides, the teaching of listening comprehension among primary pupils is often neglected even though the role of listening comprehension in language teaching has been repeatedly emphasized. Using an action research design, this study explores the teaching of listening comprehension utilizing YouTube and Video-Podcast involving 40 Primary Year 4 pupils from one of the cluster schools of excellence in the northern region of Malaysia. This study attempts to investigate to what extent did the use of YouTube and Video-Podcast enhance Primary Year 4 pupils’ listening comprehension. This study adopted multiple data collection methods including pre-post tests, reflection, observation, and focused group interviews. However, the data analysis section was mostly pre and post tests data and observation. Only these data collection and analysis ...

Dynamic Adaptive Streaming over HTTP (DASH) is a recently proposed standard that offers different versions of the same media content to adapt the delivery process over the Internet to dynamic bandwidth fluctuations and different user device capabilities. The peer-to-peer (P2P) paradigm for video streaming allows us to leverage the cooperation among peers, guaranteeing the service of video requests with increased scalability and reduced cost. We propose to combine these two approaches in a P2P-DASH architecture, exploiting the potentiality of both. The new platform is made of several swarms and a different DASH representation is streamed within each of them; unlike client-server DASH architectures, where each client autonomously selects which version to download according to current network conditions and to its device resources, we put forth a new rate control strategy implemented at peer site to maintain a good viewing quality to the local user and to simultaneously guarantee the s...