main-content

## Über dieses Buch

This book constitutes the refereed proceedings of the First Workshop on Embracing Global Computing in Emerging Economies, EGC 2015, held in Almaty, Kazakhstan, in February 2015.
The 13 revised full papers presented were carefully reviewed and selected from 28 submissions. The papers focus on the topic in two ways: one part is directly relevant to the problems of delivering cloud services in an emerging economy such as Kazakhstan, the other part represents ICT innovation by scientists of the region.

## Inhaltsverzeichnis

### Building Cloud Applications for Challenged Networks

Abstract
Cloud computing has seen vast advancements and uptake in many parts of the world. However, many of the design patterns and deployment models are not very suitable for locations with challenged networks such as countries with no nearby datacenters. This paper describes the problem and discusses the options available for such locations, focusing specifically on community clouds as a short-term solution. The paper highlights the impact of recent trends in the development of cloud applications and how changing these could better help deployment in challenged networks. The paper also outlines the consequent challenges in bridging different cloud deployments, also known as cross-cloud computing.
Yehia Elkhatib

### Agent Knowledge and Beliefs in a Cloud

Abstract
Cloud computing is a concept that is in use since late 2000s related to consumption of distributed computer resources, namely servers and networks for data storage and access. In the paper we examine knowledge-based algorithms for agents that have access to a resource center to use some of available discrete resources. We assume that resource items are passive, they form a cloud, any item can be lend on demand to any agent if and only if there is no races for this item with other agents. All agents are rational and can communicate with each other in P2P-manner, negotiate, flip and swap (change intentions) so that all flips/swaps always must be rational for participating agents. The problem is to design a multiagent algorithm, which allows each agent sooner or later to access some resource item. We present a uniform algorithm scheme and then specialize for the following particular problems: Robots in Space and Rational Agents at the Marketplace.
N. V. Shilov

### Could We Fit the Internet in a Box?

Abstract
It is estimated that only 40 % of the world’s households are connected to the Internet. Half of them are in less developed countries, where household Internet penetration has only reached 28 %. This is in stark contrast to the 78 % of households in more developed countries. A key challenge facing the next generation is therefore enabling wider participation in the Internet, as well as the benefits it brings. This paper explores the feasibility of capturing network applications and services in a single locally usable “Internet Box”. The Box will operate independently from the rest of the Internet, allowing those without traditional connectivity to use the “Internet” in an simulated and disconnected manner. We conclude that the concepts have great potential, and explore some of the remaining challenges, as well as milestones achieved in the literature so far.
Gareth Tyson, Arjuna Sathiaseelan, Jörg Ott

### Exemplified Study of WRED Algorithm Parameters Influence on IP-based Networks

Abstract
Within the TCP/IP model, if the transport layer of multiple machines send too many packets at high speed, the network will be quickly overloaded and productivity of the system will sharply fall causing delay and packet loss. Congestion control, aimed at combating such situations requires collaboration between the network and transport layers. Since overload occurs on routers, their discovery has been on the side of network layer of TCP/IP protocol stack. However the reason for congestion is the traffic transmitted by the transport layer. Therefore, the only effective way to control congestion is slower packet transmission by transport protocols and algorithms. Weighted random early detection (WRED) is a queuing discipline for a network scheduler suited for congestion avoidance. It has an advantage in comparison with other congestion control techniques, but at the same time requires very correct settings. This paper examines the algorithm of the selecting the correct setting for the WRED algorithm on the example of the widespread corporate networks.
Kuat Abdrau

### Mobility Types for Cloud Computing

Abstract
We propose a mobility type system for description and verification of distributed systems in which processes are asked to move between locations where important local interactions are required. We use a simple version of distributed $$\pi$$-calculus to define mobility types. The novelty of this approach is that we point out sequences of migrations as global types, and investigate scenarios in which processes are required to follow such a sequence of migrations along several locations. The typing system ensures certain properties including type soundness.
Bogdan Aman, Gabriel Ciobanu

### Reversible Express Bus Lanes Simulation Software

Abstract
Exempting buses from other traffic seems to be an obviously good idea. On the other hand, it is not always possible to dedicate necessary space in the city for an extra lane. Therefore, our goal was to optimize the existing roads for giving priority to buses. Again, it would be nice to dedicate the rightmost lane to buses, but the other drivers would feel strong discomfort when turning right or parking. Moreover, this approach requires two lanes, one lane in each direction. So we tried to examine whether it makes sense to take the lane in the median area of the road, make it reversible and give it to buses.
In our work we simulated such a reversible bus lane. We simulate buses that have to change to the leftmost lane and then change back to the rightmost lane for a bus stop. This way we have found out the minimum distance between bus stops such that buses would be able to use it.
Timur Bakibayev, Gulnara Bekmagambetova, Asem Turarbek

### A Combinatorial Approach to Knot Recognition

Abstract
This is a report on our ongoing research on a combinatorial approach to knot recognition, using coloring of knots by certain algebraic objects called quandles. The aim of the paper is to summarize the mathematical theory of knot coloring in a compact, accessible manner, and to show how to use it for computational purposes. In particular, we address how to determine colorability of a knot, and propose to use SAT solving to search for colorings. The computational complexity of the problem, both in theory and in our implementation, is discussed. In the last part, we explain how coloring can be utilized in knot recognition.
Andrew Fish, Alexei Lisitsa, David Stanovský

### MultiBox: Lightweight Containers for Vendor-Independent Multi-cloud Deployments

Abstract
Cloud computing aims to make a large selection of sophisticated technologies available to users for deployment and migration. In reality, once a cloud service provider has been chosen, migration is often a costly and time-consuming process. This paper presents MultiBox, a lightweight container technology that facilitates flexible vendor-independent migration. Our framework allows its users to deploy and migrate almost any application in its normal state with minimal computational and network resource overheads. We show that the performance overhead of deploying within a lightweight container is 4.90 % of the resources available to an average VM and downtime during a migration is less than the time needed to scale a server using provider-centric tools.
James Hadley, Yehia Elkhatib, Gordon Blair, Utz Roedig

### Texture Recognition by Spatially Adaptive Classification

Abstract
The image preprocessing and the skeleton orientation method are applied to segment a texture image with structure-oriented patterns. The technique is incorporated with a spatially adaptive classification of geometric features. The algorithm is tested on a set of artificial images and X-ray tomography scan of titanium alloy. The results are presented and discussed.
Anatoly Kornev

### A Unifying Framework for Interactive Programming and Applications to Communicating Peer-to-Peer Systems

Abstract
We present a unifying framework for interactive programming built-up on top of rv-systems, a space-time invariant model for interactive computation which extends register machines with interactive capabilities. The framework provides a kernel coordination language parametrized by the programming language used for writing the code of the basic blocks. To illustrate the framework, we describe an implementation of an interactive communicating system where the nodes use Chord peer-to-peer communication protocol and their coordination is written in Agapia, a structured interactive programming language for rv-systems.
Alexandru Popa, Iulia Teodora Banu-Demergian, Camelia Chira, Florian Mircea Boian, Gheorghe Stefanescu

### Multiscale Blood Vessel Segmentation in Retinal Fundus Images Algorithm Implementation and Analysis

Abstract
Image segmentation helps to analyze images by simplifying the representation of image. It is clear that there is no universal algorithm for image segmentation methods; different methods should be used depending on the application. In this paper multiscale blood vessel segmentation in retinal fundus images algorithm [1] was implemented and its parts were analyzed. In order to reduce noise, OpenCV blurring functions were used. Moreover, the problem of segmentation was described. It was observed that the blood vessel can be identified using the multiscale blood vessel segmentation in retinal fundus images algorithm. It also found that the preprocessing of the captured fundus images is very essential. Thus the results can be further enhanced by using selective and regional image smoothing functions according to the fundus images characteristics before applying the multiscale blood vessel segmentation in retinal fundus images algorithm.
Aigerim Sarbasova, Md. Mahmud Hasan

### Generating Mediators for Dynamic Interoperability Across Cloud Providers

Abstract
With the growing popularity of cloud computing and large number of providers, it is necessary to select the most appropriate service. To select a suitable provider the user needs to browse a huge list of providers. To solve this problem, the customer can use a third-party service called a cloud broker. One of the problem we faced during the developing cloud broker framework is a lack of API standards. Every provider has its own protocol for deploying services. This paper describes a new approach to construct a protocol for communication between the client and the provider. It is based on the use of mediators, which, depending on the selected service capable to convert messages from a format of one particular provider to a format of another provider. All the processes described in a specific example of Google Compute Engine client and Digital Ocean server interactions for deploying the service. In addition, during the development of mediators was introduced a new operator “sync” to synchronize the set of answers to one.
Oleg Trubitsin

### Verification and Validation of Formal Data-Centric Business Models

Abstract
This paper addresses the problem of describing and analysing internally consistent data within business process workflow specifications. We use Rodin platform for verifying the correctness of the Event-B models. These models we obtain from an ontology and an associated set of normative rules by applying mapping rules. The latter enable us to transform these specifications into Event-B modular artefacts. The resulting model, by virtue of the Event-B formalism, is very close to a typical loosely coupled component-based implementation of a business system workflow, but has the additional value of being amenable to theorem proving techniques to check and refine data representation with respect to process evolution. In this paper, we give a formal account of the design specifications defined by Event-B modules and perform verification and validation by using theorem proving techniques provided by Rodin platform.
Timur Umarov, Rustem Kamun, Askhat Omarov, Sanzhar Altayev

### Backmatter

Weitere Informationen