Interoperability and automatic adaptation for novel generation middleware systems

November 26th - 27th, 2015

Dr. Etienne Rivière, University of Neuchâtel

The following four professors will give the lectures for this event:

• Prof Gordon Blair, Lancaster University, UK - Prof Blair is the author of the authoritative textbook on distributed systems
• Prof Laurent Réveillère, University of Bordeaux, France
• Prof François Taiani, Université of Rennes, France
• Prof Virgil Dobrota, Technical University of Cluj-Napoca, Romania

In addition, participating students will be offered to give short presentation of their works and get feedback from other participants.

Middleware systems allow abstracting the complexity and diversity of large-scale and heterogeneous distributed systems. Recently developed tools and techniques are changing the way we consider and operate such systems. Getting to know these novel techniques is important for PhD students working in the field of distributed systems at large, but will also benefit to students interested in languages, software engineering, automatic adaptation and runtimes, among other fields.

This doctoral school will cover middleware construction and operation for complex systems comprising large numbers of heterogeneous and geographically distributed entities. Starting from an introduction to the field and a perspective on the challenges and achievements (Prof Blair), lectures will focus on tackling heterogeneity and adaptation (Prof Réveillère), using self-organizing techniques to build complex large-scale systems (Prof Taiani) and using programmable networks (Prof Dobrota).

UniNE - room B013 - see http://www2.unine.ch/iiun/page-5149_en.html for travel instructions

Map

Tentative agenda follows. The doctoral school will take place in all cases from November 26 9am to November 27 12:15am.

Thursday November 269:00 opening
9:15-11:15 Lecture 1: Prof Gordon Blair
11:15-11:30 break
11:30-12:15 students presentations
12:30-14:00 lunch (on site)
14:00-16:00 Lecture 2: Prof Laurent Réveillère
16:00-16:15 break
16:15-18:15 Lecture 3: Prof François Taiani

Friday November 27
9:00 opening
9:15-11:15 Lecture 4: Prof Virgil Dobrota
11:15-11:30 break
11:30-12:15 students presentations (x2-3)

The abstracts of the lectures are provided below.

Prof. Blair: "Middleware for Complex Distributed Systems: The Quest for New Landscapes"Distributed systems are at a watershed due to their increasing complexity. The heart of the problem is the extreme level of heterogeneity exhibited by contemporary distributed systems coupled with the need to be dynamic and responsive to change. In effect, we have moved from distributed systems to systems of systems. Following on from this, middleware is also at a watershed. The traditional view of middleware is no longer valid (i.e. as a layer of abstraction, masking the complexity of the underlying distributed system and providing a high-level programming model). In practice, middleware is often by-passed with complex systems constructed in a rather ad hoc manner as a mash-up of a variety of technologies. The end result is that middleware is no longer sure of its form or purpose and this lack of a viable approach is a huge barrier to the emergence of areas such as smart cities and emergency response systems.

This lecture will look in detail at the increasing complexity in distributed systems, highlighted with examples from the domain of Earth Observation and Management, an area where the presenter has experience of applying distributed systems technologies. The lecture will also look at the future of distributed systems/ middleware, and the presenter will argue that there is a need to fundamentally rethink the landscape of middleware in what we refer to as complex distributed systems. The lecture will examine a number of fresh perspectives on the problem, and some of the key themes discussed will be picked up in later talks. The fresh perspectives include a move to emergent middleware, seeking flexible meta-structures for distributed systems, and a step away from generic to domain-specific technologies.

Prof. Réveillère : "Interoperability for Heterogeneous Complex Distributed Systems"Nowadays, complex distributed systems are composed from systems that are developed independently of one another, including legacy systems. This composition occurs either statically, or at runtime as in the case of spontaneous interactions between mobile and pervasive systems. However, existing systems are highly heterogeneous in their interaction methods making such composition challenging. Applications and systems are developed using a multitude of incompatible middleware abstractions and protocols. The range of incompatible protocols drastically limits interoperability, and thus the practical benefit of systems composition. Protocol standardization should address this issue but has been demonstrably ineffective in practice. In this lecture, we will discuss the main challenges of interoperability and present some approaches that address heterogeneity at both the application and middleware-level, where application differences are seen in terms of incompatible interface signatures and data content, and at the middleware level in terms of heterogeneous communication protocols.

Prof. Taiani: "Large-Scale Systems And Self-Organization"Modern distributed services increasingly execute on thousands or more machines, while delivering services to millions of users. These scales are in stark contrast to that of traditional distributed systems by several orders of magnitude, and question the traditional engineering approach used to construct distributed systems. In particular, as systems continue to grow and to integrate diverse and partly autonomous subsystems, deciding beforehand on a static system architecture becomes increasingly difficult. Self-organizing mechanisms become in this case particularly useful. In this lecture, we will discuss the main challenges of self-organization in large-scale systems, and present some representative examples of epidemic-based self-organizing protocols and their application.

Prof. Dobrota: "A telecom perspective on SDN and OpenFlow"
When the provocative concept of software-defined networking was attached to the promising open-source protocol OpenFlow, the networking community expected revolutionary innovations. The cloud implementers have claimed that this new technology might be exactly what they were looking for. The major vendors tried to mix it with the legacy solutions and separate the application world from the physical infrastructure using overlay networks virtualization. But how modern is a solution that promotes a central control plane when the trend is to have a distributed approach? Will it go through the same painstaking process as IPv6? This course presents my personal view on this topic, based on a telecom perspective. Conceptual aspects such as data plane, control plane, flow tables, SDN controllers, and virtual switches will be accompanied by practical examples to help shed some light on these questions.

20