Architectural transformations in network services and distributed systems

Among them are the Internet of Services, Clouds, Smart Grids, Parallel and Distributed Computing, FogComputing and the Internet of Things, to mention a couple of popular concepts.. Reade

Andriy Luntovskyy and Josef Spillner

Architectural Transformations in Network Services and Distributed Systems

Andriy Luntovskyy

Berufsakademie Sachsen, Dresden, Germany

Josef Spillner

Service Prototyping Lab, Zürcher Hochschule für Angewandte

Wissenschaften, Winterthur, Switzerland

ISBN 978-3-658-14840-9 e-ISBN 978-3-658-14842-3

DOI 10.1007/978-3-658-14842-3

Library of Congress Control Number: 2016957988

© Springer Fachmedien Wiesbaden GmbH 2017

This work is subject to copyright All rights are reserved by the Publisher,whether the whole or part of the material is concerned, specifically the rights

of translation, reprinting, reuse of illustrations, recitation, broadcasting,

reproduction on microfilms or in any other physical way, and transmission orinformation storage and retrieval, electronic adaptation, computer software,

or by similar or dissimilar methodology now known or hereafter developed

The use of general descriptive names, registered names, trademarks, servicemarks, etc in this publication does not imply, even in the absence of a

specific statement, that such names are exempt from the relevant protectivelaws and regulations and therefore free for general use

The publisher, the authors and the editors are safe to assume that the adviceand information in this book are believed to be true and accurate at the date

of publication Neither the publisher nor the authors or the editors give awarranty, express or implied, with respect to the material contained herein orfor any errors or omissions that may have been made

Printed on acid-free paper

This Springer Vieweg imprint is published by Springer Nature

The registered company is Springer Fachmedien Wiesbaden GmbHThe registered company address is: Abraham-Lincoln-Str 46, 65189Wiesbaden, Germany

To our caring wives.

About the Book

Book objectives You are reading a book which aims to cover the broad field

of recent innovations in network services and distributed systems The targetgroup of the book encompasses students of universities and technical highschools, graduated engineers as well as teaching staff If you are somebodyelse, do not worry, the covered subjects may still be of interest to you! Thisbook offers its readers a dual functionality:

As a monograph… With the given work, we decided to help not only the

readers and students, but also ourselves, as the professionals who are activelyinvolved in the networking, telecommunications and systems communities,

by understanding the trends which have developed in the recent decade indistributed systems and networking applications Important architecturaltransformations of modern distributed systems are examined and presented insurvey style Examples of new architectural solutions for network (mobile)services and applications are discussed Among them are the Internet of

Services, Clouds, Smart Grids, Parallel and Distributed Computing, FogComputing and the Internet of Things, to mention a couple of popular

concepts

As a handbook… Current technologies, standards and research results for

advanced (mobile) networks, connected devices and provisioned services aswell as for higher-level network functions and software applications are

focused within this book from a practical angle The authors highlight howthese technical underpinnings to our digital communication and collaborationinfrastructure are being transformed to reflect society requirements Efficientarchitectures, principles and systems for mobile and wireless communication,criteria for optimisation of networks and distributed systems, as well as

central ideas to new system concepts are widely discussed herein Use casepresentations and studies with in-depth technical descriptions along with atest exam strengthen the nature of this book as handbook to use for coursesand projects

Learning objectives The learning objectives targeted by the book are as

follows:

1 Readers/students should be able to combine, integrate, analyse and

manage the solutions to the above-mentioned technologies (Clouds,

Smart Grids, Parallel and Distributed Computing, Fog Computing,

Internet of Services, Internet of Things) They should also be able to

implement custom systems on the basis of an adequate conceptual

grounding in practical projects

2 As a result, readers/students become skilled to create and evaluate performing, reliable and secure access aspects to data and network

well-applications, distributed systems and mobile apps The systems and

services should be usable in a data protection-compliant manner and

aligned with user preferences

3 Readers/students become educated to develop custom architectures ofnetwork services and distributed systems as well as to comment critically

on the associated problems

Numerous examples in the chapters, comparison tables, excursions intotechnological stacks, figures with structures and demonstrations are

highlights of this book Every chapter has a list of keywords, complemented

by actual system examples, a summary and continuing bibliographic records.Furthermore, at the end, there is a whole chapter dedicated to repetition andself-controlling by offering questions and answers to many of the discussedtopics along with further insight into the research behind the covered systemsand services

Motivation Despite the existence of a broad range of scientific and

practical literature on the topics of distributed computing, cloud computing,privacy-preserving systems, green IT, Internet of Things and so forth, fromour perspective as researchers and lecturers there is a distinct lack of

combined monographs/handbooks with a pretence to be useful to education

In particular, most of the literature describes technological snapshots as

points in time Instead, we want to explicitly include historical backgroundinformation and focus on the ongoing evolution and trends which are similar

in many areas Furthermore, we were not satisfied with literature which

merely lists positions and standards instead of allowing the reader to diveright into the technology by offering concrete implementation and use case

links Especially for students in co-education (for instance

Berufsakademie/Duale Hochschule , Fachhochschule and (höhere)

Fachschule in Germany and Switzerland, vocational and community colleges

in the U.S.), the practical links are essential to decide whether or not a certaintechnology should indeed be evaluated for upcoming projects

The book partially continues the educational approach of a previous book

called Planning and Optimisation of Computer Networks: Methods, Models,

Tools for Design, Diagnosis and Management in the Lifecycle of Wired and Wireless Computer Networks , by Luntovskyy, Guetter and Melnyk, which

appeared by Springer Vieweg in German language in 2011 The original title

is: Andriy Luntovskyy, Dietbert Guetter, Igor Melnyk Planung und

Optimierung von Rechnernetzen: Methoden, Modelle, Tools für Entwurf, Diagnose und Management im Lebenszyklus von drahtgebundenen und

drahtlosen Rechnernetzen Springer Fachmedien Wiesbaden GmbH, 2011,

435 pages (ISBN 978-3-8348-1458-6), 1st edition 2011, with 245 figures und

64 tables The present book complements and extends the range of topics Itaddresses the evolved development from computer networks to network-integrated and network-connected services, in particular cloud and fog

services, as well as modern architectures of distributed (mobile) applicationssuch as 5G and low-energy radio links The new book therefore presents aholistic view on transformation processes, which are nowadays often lesstechnically motivated, but rather by the needs of the society which is subject

to a higher degree of pervasive services The benefits for society are aboutecology (green networks), privacy (secure clouds), comfort (always on) andeconomy (pay as you go)

Structure of the Book

This book is divided into seven chapters The first chapter offers a bird’sperspective on the history and present development of networking and

service topics The second chapter presents state-of-the-art distributed

systems and uses them to explain the architectural transformations whichmost of today’s systems are subject to In the chapters three to six, differentarchitectures and systems will be presented, including clusters, clouds, fogsand mobile applications The seventh chapter offers a holistic view on

security in networked services Finally, five appendices and one more

auxiliar digital appendix complete the book

Chapter 1 – Periodisation of Network Service Development The

evolution of hardware and infrastructure on one hand and of services onthe other hand is divided into four phases each

Chapter 2 – Architectural Transformation in Distributed Systems

Clusters and clouds, peer-to-peer architectures and distributed databaseswill be presented and reflected on in the context of the evolution andtransformation of systems

Chapter 3 – Evolution of Clustering and Parallel Computing Clusters,grids and parallel computing will be introduced Their benefits

concerning the performance of computing, but also the necessary offs with energy consumption and price, will be highlighted The

trade-management of resources and applications in these environments willalso be explained

Chapter 4 – Cloud Computing, Virtualisation, RAICs and SDN Thischapter will introduce contemporary cloud stacks and services, includingprogrammable networks, virtual teleconferences and safe data backups.Chapter 5 – Smart Grid, Internet of Things and Fog Computing Beyondthe software side, small connected hardware devices and the connectionbetween computer networks and energy distribution networks will becovered in this chapter

Chapter 6 – Future Mobile Communications: From 4G to 5G, 5G

Enabling Techniques Mobile communication protocols for global

(phones) and local distances will be presented A special focus is on theupcoming 5G connectivity

Chapter 7 – Security in Distributed Systems This chapter will give aholistic view on what is commonly called security, by introducing intoconcrete protection goals and matching security layers It will also

include a discussion of privacy and legal aspects with a focus on howusers can protect their activities and communication in today’s and

tomorrow’s distributed systems

Appendices First, selected originators and designers of distributed

systems will be briefly presented Then, specific research projects withrecent results which contribute to the evolution and transformation will

be introduced The further parts contain explanations to common

acronyms in mobile and wireless technologies, a repetition and controlpart to track the learning progress when reading the book, and finally anexample of a written exam to the discussed subjects The solutions to theexam are available as auxiliar digital appendix.

Andriy Luntovskyy

Josef Spillner Dresden, Germany, Winterthur, Switzerland

About the Authors

The book contents have been primarily provided by Andriy Luntovskyy.Some sections and editorial guidance were provided by Josef Spillner Most

of the material is published for the first time, although some is based onprevious research papers, including joint papers by the authors, and materialkindly added by fellow academics

Andriy Luntovskyy, Prof Dr habil.

Andriy Luntovskyy is with BA Dresden University of Cooperative

Education, Dresden, Germany

Office: Room 2.105., Hans-Grundig-Strasse 25, 01307 Dresden

Andriy Luntovskyy is member of the Academy of Sciences for HighSchool of Ukraine (ANVSU.org.ua) and member of the Academy of

Telecommunications of Ukraine and International IT Academy

Teaching and Classes: Computer Networks, Mobile Communication andTelematics, Basics of Programming and Software Technology, DistributedSystems, Operating Systems, Web-Applications and Office Communication,Data Security and IT Legacy, Basics of Computer Science and Business

Informatics Guest lectures in Ukraine and Poland, classes for bachelor,

master and PhD students

Research: CANDY – Computer-Aided Network Design utility, Design ofWired, Wireless and Mobile Networks, Clouds, Clustering and Mobile

Computing, Web Services, SOA and Virtualisation Methods, Mobile andWireless Networks, Energy Efficiency in Networks, Wireless Sensor,

Networks, Smart Grid and IoT, Multiservice Mobile Platforms

Attendance and co-chairman at multiple conferences and forums (CEBIT

2007, 2008, 2011) Publications: two books are published in Germany (2008,2011); other 12 books in mother tongue in Ukraine; more than 130 papers toconferences and magazines, among them multiple IEEE Xplore publications

Josef Spillner, Dozent, Dr.-Ing habil.

Josef Spillner is with Zurich University of Applied Sciences (ZHAW),School of Engineering, Winterthur, Switzerland

Office: Room O3.17., Obere Kirchgasse 2, 8400 Winterthur, Switzerland.Phone: +41 (0) 58 934 45 82

Teaching and classes: Introduction into Research Areas of ComputerScience, Development of Distributed Sysstems on the Basis of SOA,

Complex Internship for Service and Cloud Computing, OS and ComputerNetworks, Basics of Programming and Software Technology, Distributed

Systems, Python Programming Classes for bachelor and master students aswell as non-IT students, in particular media informatics and industrial

engineers

Research: THESEUS/TEXO – New Technologies for the Internet ofServices, funded by the German Ministry of Economics (BMWi) FlexCloud– Flexible Service Architectures for Cloud Computing, funded by the

European Social Fund (ESF) DaaMob – Service-oriented Platform Conceptsfor Cross-System Third-Party Applications with Mobile Components in theInternet of Things, funded by the German Research Council (DFG) Furtherresearch on XML Schema, Web Service GUIs, Cloud Controllers, CloudCockpits and Energy Efficiency, Stealth Computing

Attendance and involvement with multiple conferences and workshops.Publications: books co-authorship, more than 50 papers and journal articles,technical reports with HPI Future SOC Lab, IEEE and ACM conferencechairing

All our graceful heart’s acknowledgements to Prof Dr rer nat habil Dr h

c Alexander Schill (encouragements and challenges), Dr rer nat DietbertGütter (proofreading), Prof Dr Andreas Westfeld, Prof Dr Thomas Horn,

Dr Reiner Keil (inspiration, in absentia) and many other colleagues, studentsand reviewers for their helpful and friendly support, the inspirations and co-operation while completing this work

Our special acknowledgment goes to Dr.-Ing habil Igor Melnyk for hisaltruistic contribution to the modelling of the waste heat and cooling process

in “green” data centers and clouds

Distributed Denial of Service 2

Distributed Input Distributed Output 211, 225, 230, 241–244

Domain Name System 15

Digital Subscriber Line 2, 3

Enterprise Application Integration 81, 82, 84, 98

Electro-Magnetic 138, 139, 141, 168, 174, 208

Energy Reuse Efficiency 136

Enterprise Service Bus 16

Floating-Point Operations Per Second 46–50, 54, 59–61, 67File System in Userspace 124

Global System for Mobile Communications 211, 212

High-Performance Computing 45, 59, 66

High Speed Download Packet Access 211, 213, 217

Hyper-Text Transport Protocol 80, 82–84, 90

Heating, Ventilating and Air Conditioning 8, 9

Infrastructure-as-a-Service 9, 77, 79, 81, 85

Internet Control Message Protocol 8, 140

Intrusion Detection System 247, 270, 271, 273, 274, 276

Internet Engineering Task Force 6

Intrusion Prevention System 247, 270–273, 276

Integrated Services Digital Network 1, 6

KNX Home and Building Control Standard 7, 9, 140

Local Area Network 8, 19, 140, 159, 174, 176, 178, 195, 198, 201Low-Energy Adaptive Clustering Hierarchy 166

Local Operating Network 7, 9, 140

Open Services Gateway Initiative 15

Open Systems Interconnect 144, 145, 259, 261, 272

Peer-to-Peer 13, 19–23, 43

Platform-as-a-Service 9, 77, 81, 85, 86, 89

Primary-Copy-Schema 37

Plug-in (Hybrid) Electric Vehicles 138, 140, 141, 149

Pretty Good Privacy 247, 260, 262, 276, 277, 289

Power Line Communication 135, 148, 158, 159, 189

Power over Ethernet 9

Power Usage Effectiveness 3, 9, 136, 150, 151, 153–155, 157, 158Quality of Experience 85, 86

Quality of Service 1–4, 17, 77, 79, 82, 85, 86, 110, 113, 114, 138,

154, 161, 166, 197, 208

Redundant Array of Independent Clouds 77, 91, 111, 113, 119–

123, 125–131

Redundant Array of Independent Disks 112, 113, 119–122

Representational State Transfer 82–85, 89

Requests for Comments 6, 213, 214

Rivest, Shamir, Adleman Cryptosystem 255, 257, 258, 265, 276,

282, 283

Software-as-a-Service 9, 77, 79, 81, 85, 89, 92

Storage-Area Network 81, 112, 113

Software-Defined Networking 77, 92, 105–110, 225, 230, 232Secure Electronic Transaction 279, 281, 283, 284, 287, 288

Stateful Inspection Firewall 247, 270–272

Session Initiation Protocol 213, 214

Service Level Agreement 2, 81, 82, 85, 91

Small and Medium Enterprise 7, 139

Stateful Multi-Layer Inspection Firewall 247, 272, 276

Symmetric Multi-Processing 60, 61

Simple Mail Transmission Protocol 6

Simple Network Management Protocol 8, 140, 192

Service-Oriented Architecture 79, 82–84, 88, 89, 98, 113

Simple Object Access Protocol 83, 85, 90

Structured Query Language 30, 35, 36, 39, 40

Secure Sockets Layer 264–266

Transport-Layer Security 247, 256, 258, 260, 264, 265, 267, 268,

279, 281–283, 287, 288, 303

Univeral Mobile Telecommunications System 211, 213, 214

Universal Plug and Play 15

Virtual Machine 85, 92–95, 98, 100, 105–108, 114

Voice over IP 214, 215

Virtual Private Network 247, 265–268, 270, 271, 281, 283, 302Virtual Telecommunication Engineering Offices 77, 84, 85, 88–91World Wide Web Consortium 6

Web Application Firewall 247, 270, 276

Wireless Area Network 145, 159

Worldwide Interoperability for Microwave Access 139, 149,159–161, 178, 188

Wireless Local Area Network 8, 19, 140, 159, 161, 171, 178, 180,

187, 195, 202, 211, 212, 214, 224, 225, 234, 240, 241, 243, 244

Wireless Personal Area Network 19, 135, 158, 168

Wireless Sensor Networks 139, 141, 161–166, 173, 174, 189

Everything-as-a-Service 79

Extensible Messaging and Presence Protocol 70, 73, 74, 85

1 Periodisation of Network Service Development

References

2 Architectural Transformations in Distributed Systems

2.1 Software Architectures and Communication Patterns

2.2 Distributed Service Systems: Clustering, Grids and Clouds

3 Evolution of Clustering and Parallel Computing

3.1 Clustering and Grids: Performance Parameters and Basic Models 3.2 Performance-Energy-Price Trade-Offs in Clusters and Grids 3.3 Resource Management in Clusters

3.4 Application Management in Clusters

3.5 Application Management in Grids

3.6 Distributed Applications

3.7 Conclusions

References

4 Cloud Computing, Virtualisation, Storage and Networking

4.1 Clouds: Technology Stack, Basic Models and Services

4.2 Virtualisation of Services and Resources

4.5 RAIC Integration for Network Storages on Mobile Devices

4.5.1 Efficient Access to Storage Services from Mobile Devices 4.5.2 A New Must-Have App: RAIC Integrator for Smartphones 4.6 Conclusions

References

5 Smart Grid, Internet of Things and Fog Computing

5.1 Smart Grid as Integration Technology for the Networks of

Energy Supply and Telecommunication

5.1.1 Services Architectures and Multi-level Models

5.1.2 Smart Grid: Enabling Network Technologies

5.1.3 Case Study: A CAD Toolset for the Design of Efficient Combined Networks

Energy-5.2 From Internet of Services to Internet of Things: Fog Computing 5.2.1 Enabling Technologies for IoT

5.2.2 Case Studies on IoT with On-Board Micro-controller

6.1.2 Satellite-Based Radio Systems

6.2 A New Generation of Mobile Communication

6.2.1 Visions and Requirements

6.2.2 5G Inter-Operability

6.2.3 Future Standard IMT 2020 Deployment Scenarios

6.2.4 Resource Allocation Method for Future WLAN

6.3 Conclusions

7 Security in Distributed Systems

7.1 Security and Protection Goals

7.3.3 Content Encryption: S/MIME and PGP

7.3.4 Authorisation: Kerberos and OAuth2

7.3.5 Further Secure Services: DNS-SEC, VPNs and Proxies 7.4 Security Protocols and Network Concepts

7.5 Firewalls

7.6 Security in Web Applications: Legal and Technological Aspects 7.6.1 Technological Aspects of Data Security Guaranteeing Web Systems

7.6.2 Legal Aspects of Data Security Guaranteeing Web Systems 7.7 Steganography in Distributed Systems

7.7.1 Steganography in Development

7.7.2 Steganography: Main Concepts

7.7.3 Watermarks and Steganography

7.8 Anonymity and MIX Networks

7.9 Conclusions

References

Appendix A Selected Originators and Designers of Distributed Systems A.1 Edgar Frank “Ted” Codd

A.2 Tom DeMarco

A.3 Grady Booch

A.4 James Gosling

A.5 Sir Timothy John Berners-Lee

A.6 Tim O‘Reilly

A.7 Roy Thomas Fielding

A.8 Sergey Brin

A.9 Philip R Zimmermann

A.10 Remembering the Pioneers

Appendix B Research Focus

B.1 CANDY: Network Planning

B.2 FlexCloud: Flexible Architectures for Cloud Computing

B.3 DaaMob Service Platform: Data Service Management

Appendix C Acronyms for Mobile and Wireless

Appendix D Repetition and Control of Learning Progress

D.1 New Generation (Mobile) Networks

D.2 Periodisation of Computer Networks: Phases I to IV, Smart Grid, IoT and Fog Computing

D.3 Architectural Transformation in Distributed Systems

(2)

© Springer Fachmedien Wiesbaden GmbH 2017

Andriy Luntovskyy and Josef Spillner, Architectural Transformations in Network Services and

Distributed Systems, DOI 10.1007/978-3-658-14842-3_1

1 Periodisation of Network Service

Development

Andriy Luntovskyy1 and Josef Spillner2

Berufsakademie Sachsen, Dresden, Germany

Service Prototyping Lab, Zürcher Hochschule für Angewandte

Wissenschaften, Winterthur, Switzerland

Andriy Luntovskyy (Corresponding author)

Email: andriy.luntovskyy@ba-dresden.de

Josef Spillner

Email: josef.spillner@zhaw.ch

Keywords Networks – Services – Quality of Service ( → QoS) – Internet of

Services ( → IoS) – Clouds – Smart grid – Internet of Things ( → IoT) – Fog

computing

Information and communication technology is moving fast What are gridsfor nowadays? Is anybody still using Integrated Services Digital Network ( 

→ ISDN) connections? Will the ‘digital fog’ be around all of our devices, and

for how long when on batteries? What is the cost of safely storing one digitalphoto taken on the mobile phone for the rest of our lifetime? Readers whohave immediate answers to such questions are asked to put this book asideand spend their time with more pleasure All other readers are however

invited to follow us briefly through the history of network services and

distributed systems, through the past transformations and current trends, in

order to learn about the rather complex landscape of distributed service

systems in the future These digital, physical and combined (cyber-physical)systems affect our daily lives, as we interact with them through screens anddevices, software applications, processes and ambient sensors

Technology development in four phases Network services and

distributed systems are two pillars of the same trend: To make applicationfunctionality provided from single computers or millions of connected

devices available to billions of people Internet and web applications

including online social networks and digital telephony already today need toscale to billions of users, which would be impossible on a single machine.Instead, many computers are clustered and many clusters are geographicallydispersed and connected so that users perceive them as single service Theperception is trained for high performance, high reliability, high privacy andsecurity, low cost, low effort and low energy consumption, among otherfactors Services not offering all of these benefits will have decreasing

chances to compete for users, and will ultimately fail to be sustainable Trustand reputation would in such cases be hard to recover

It took computer scientists and the IT industry many years to achieve thebreakthrough towards this vision In the course of development of networkedapplications and services, including telecommunication, web and cloud

services offered on-demand in any situation, four distinct phases in the

technological foundation can be identified

The first phase starting with the roll-out of networks and the Internet(about 1970–2000) had the purpose of offering the functionality and of

ensuring improvements to the QoS The QoS considerations were mostly

confined to strict technical network characteristics, without taking end-to-enduser experience into account Bandwidth increased and latency decreased Toput the bandwidth development into perspective: In 1999, a 56 kbit/s modemconnected to copper telephony networks was the norm for private users and

just about to be replaced by faster Digital Subscriber Line ( → DSL)

connections with about 768 kbit/s downstream bandwidth Consumers couldonly rely on such numbers as upper bounds in a best-effort service marketand could not easily translate these numbers into application benefits, forinstance, video quality or file transfer performance

In the enterprise market, large computing centers were economicallyeffective due to using broadband Internet connections which enabled theconsolidation of a lot of compute and storage resources behind a single data

pipe They helped also in mitigation of Distributed Denial of Service ( 

→ DDoS) attacks due to load distribution between several servers and links.

The system reliability was improved due to better availability of spare parts(hard drives, power units, switches etc.), the employment of redundant unitswherever possible, and emergency power generators in large centers, wherethey were feasible Similarly, the application availability and scalability wasincreased with replicated setups in high-availability/failover and load-

balancer setups, respectively

Ultimately, the phase has been about connecting people to the Internet, inother words, an Internet of People A simple formula characterises the firstphase:

(1.1)

In the second phase of development of Internet services (about 2000–

2010), the improvement of QoS was accompanied by explicit cost

optimisation, among other reasons due to hardware consolidation and server

virtualisation in combination with QoS guarantees codified in a Service Level Agreement ( → SLA) These mandated a minimum cost by strictly given QoS

constraints But also the large size of computing centers still led indirectly toless cost on the side of customers due to the economy of scale when buyinglarge charges of spare parts and electricity The maintenance cost in the largecomputing centers is also less than in smaller ones, because the servers areupdated centrally with security patches, upgrades can be better tested beforedeploying and the maintenance actions are mostly the same at homogeneousservers To give an example: The e-commerce seller Amazon had a revenue

of about seven billion US$ in 2004 The capacity needed to operate this

business at that time is nowadays added daily to their computing

infrastructure It is not yet clear how to compare the technical characteristics

of data centres, but just looking at their dimensions demonstrates the trendtowards consolidation The Lakeside Technology Center in Chicago, one ofthe largest multi-tenant centres, has a usable surface of more than 100,000 m2across several floors of a historic printing house Microsoft’s Dublin datacentre is roughly half this size [10] Major service operators have expandedvastly during the second phase and now operate multiple of such large datacentres

On the network side, in 2009, 16 Mbps/s ADSL connection were widely

available in many urban areas in developed countries and even 55 Mbps/s

VDSL2 connections were available in selected areas, whereas in 2014,

vectoring-based VDSL brought up to 100 Mbps/s downstream and 40 Mbps/s

upstream bandwidth to consumers A slow-down in connection speed growthbecomes evident Furthermore, the promise of many governments during thistime to achieve 100 % broadband coverage had (and still has) not been

achieved anywhere Enhancing the role of hosted applications (in so-calledclouds) as integration path and cost reduction driver for applications andcomputing power characterises this second development phase

Consequently, an Internet of Services (IoS), in particular cloud services,

characterises the second phase:

(1.2)The third phase (after 2010) was triggered by the trend of “green” IT andincreasing energy demand and prices The computing centers were built moreoften in colder regions of the earth More energy-efficient hardware was

installed, and software was written with energy efficiency in mind

Processors gained dynamic voltage and frequency settings, among other

techniques, which helps shrinking the power consumption over all idle

periods The metric Power Usage Effectiveness ( → PUE) has gained

prominence, and consumers are increasingly aware and demanding of

sustainable IT The use of mobile phones to host applications and even

mobile services strengthens the awareness due to limited handset batterycapacity Smart grids installations are on the rise and lead to greater energyautonomy by turning consumers into providers Therefore, to characterise thethird phase in a formula:

(1.3)

As a by-product of the awareness, similar to transportation companieswhich can also be viewed as a public utility, the first data centre and hostingbusinesses have announced to have met a 100 % renewable energy goal [3].This has led to a voluntary green energy market which in the USA alone hasaround five million customers who have purchased, directly or indirectly,approximately 74 million MWh of power generated from renewable sources[6] In Switzerland, around 10 % of all power consumption is linked to thevarious forms of IT, an equivalent of 400,000 cars in terms of fossil fuel, and

an increasing number of providers advertise their decision to contract 100 %renewables [2]

Finally, the fourth and last phase which has already started but will cause

a high impact on computing in the near future needs to be discussed

Therefore, this book is dedicated to this phase without dismissing the earlierones Figure 1.1 puts all three already identified phases with the not yet

covered last one into context

Fig 1.1 Periodisation of network service development

The fourth phase, the next development vector, is about to happen now Thisphase is oriented not just at networking services and distributed software

applications, but to a truly user-focused IoS in many domains It happens across clouds, in the frame of the IoT with many connected small (sometimes

wearable) devices, cyber-physical systems and robots, next-generation

mobile networks and ultimately fog and wearable computing This

combination expands the always-on, always-available, pay-as-you-go utilityand cloud computing paradigm with intelligent network nodes (e.g radionetwork edges, smart routers or even smart watches) and enables via thisextension a set of new applications and services The features of such aninterpretation of fourth-phase computing are as follows:

low-latency, location-aware, energy-efficient use of heterogeneoushardware from large-scale computing centres to tiny nodes;

very big number of hardware nodes and their mobility, based on IPv6connectivity;

wide geographical distribution of miniaturised hardware, self-updatingsoftware and large volumes of data;

leading role of wireless access to connect nodes and users even overlonger distances;

service interfaces, streaming and real-time applications with guaranteed

QoS properties.

A wider interpretation of fog computing offers the appropriate platforms

for IoT, clouds and the smart grid (Fig. 1.2).

Fig 1.2 Fog computing vision (background photo: Claudia Jacquemin, JOTT Fotografie Dresden, the

depicted place: CAD/CAM system at BA Dresden – University of Cooperative Education)

According to Eric Schmidt, at that time CEO at Google, at the World

Economic Forum in Davos, Switzerland in 2015: “I will answer very simply

that the Internet will disappear… There will be so many Internet Protocol ( 

→ IP) addresses… so many devices, sensors, things that you are wearing,

things that you are interacting with that you won’t even sense it … It will be

part of your presence all the time Imagine you walk into a room, and theroom is dynamic And with your permission and all of that, you are

interacting with the things going on in the room… A highly personalised,

highly interactive and very, very interesting world emerges.” [7]

This industrial development is bound to happen, as so far the

miniaturisation of hardware is still advancing rapidly On the other hand,researchers also look into ways to keep the user in the loop and ultimatelyalso in control, something typically neglected by industrial development.Therefore, new methods for informational self-determination and

manageability of personal devices and services need to be found A typicalexample is a safe networking kill-switch to prevent any communication from

a device, something found only occasionally on devices despite its usefulnessalong with a definite off-switch Before going into the details about the futuredevelopment, the same four phases shall be analysed from a service

perspective

Network services in four phases Along with the technical

improvements in servers, devices and connectivity, the offered services

themselves have evolved over time One difference when compared to thehardware technology is the fact that new services almost always complementexisting ones instead of replacing them While it would be hard to order an

ISDN connection or a Fiber Distributed Data Interface (FDDI) connection

nowadays, we still communicate via decades-old e-mail protocols and locateservices via another decades-old domain naming protocol!

In the first phase (1970–2000), basic network services and early webapplications were created Many network services were and indeed still aredefined by an international community called the Internet Engineering Task

Force ( → IETF) in public and well-edited Requests for Comments ( → RFC)

[9] An example would be an e-mail sending service (Simple Mail

Transmission Protocol ( → SMTP)) first defined in RFC 821 by Jonathan B Postel in 1982, and subsequently updated to RFC 2821 in 2001 and finally

RFC 5321 in 2008 Other examples include real-time messaging, file transfer

and authentication Early web applications include e-commerce shops alongwith search engines and online newspapers, for instance, books.com in 1992,yahoo.com and spiegel.de in 1994, amazon.com and nytimes.com in 1995and google.com in 1997/1998 Their growth in popularity was mainly driven

by the first web browsers as client applications, including Mosaic (1992),

Netscape Navigator, Microsoft Internet Explorer and Opera (all around

1994)

The first phase also contained the first monopolisation tendencies

Whereas previously, network protocols were defined and then implemented

by multiple vendors, especially web applications emerged whose interactionwas neither well-known nor easily reimplementable Web pages as

interaction part of web applications were standardised by another entity, the

World Wide Web Consortium ( → W3C), but filled with vendor-specific

extensions which even today still cause trouble and processing overhead

In the second phase (2000–2010), due to faster home connection speeds,peer-to-peer filesharing applications became popular between consumers Anearly example has been napster.com which ceased to exist in the year 2000,only to be replaced by open protocols including Bittorrent from 2001 on.Other peer-to-peer applications quickly gained popularity, including videoconferences and in the year 2009 the cryptocurrency Bitcoin Interestingly,some applications such as permanent file storage have mostly remained withcentralised data centres, despite peer-to-peer applications being available [1].Web applications were further growing by faster and more powerful webbrowsers which emerged after a perceived innovation poise The browserswere Apple Safari (2002), Mozilla Firefox (2004) and Google Chrome (2008)which turned increasingly into a platform with all of the associated lock-inand vulnerability issues

In the third phase (since 2010), commercial global-scale services havebeen competing for marketshare Online social networking services like

facebook.com and twitter.com claim hundreds of millions of active userswhich are handled by a global network of distributed data centres Millions ofdevices and sensors are connected to enable more services And computinginfrastructure services with compute, storage and networking services haveemerged in multiple forms and concentrate applications and services in

shared data centres During this time, consumers have become increasinglyaware of where services are hosted and how they are delivered In particular,privacy issues have emerged and are not solved yet [5] Figure 1.3 contains ascheme of today’s distributed networks and services and how consumersinteract through and with them

Fig 1.3 Scheme of services and supporting hardware technology for a single distributed application

Now we can only speculate which novel services will be enabled by the

current wave of technological development This will depend in large part onthe knowledge, skills and facilities to enact new services by individual

developers and businesses The following three fictive scenarios illustrate thehypothesis about the advancement of technological trends in the fourth phase

of the chosen periodisation They will be picked up in the next chapters andillustrated with concrete examples

Scenario 1 Smart grid in an SME What will be a middle-class network

connection for an Small and Medium Enterprise ( → SME) in 2020? Only one

cable or wireless link will provide the utility services such as electricity,

telephony, Internet, digital high-definition television and cloud services

Room heating will be realised via derivation and recycling of redundant

energy from multiple (virtual) servers The wired and wireless automation of

local-area as well as piconets like Local Operating Network ( → LON), KNX Home and Building Control Standard ( → KNX), ZigBee, EnOcean will be

used to serve and control the in-door climate Management of such integratednetworks can be performed through Ethernet Local Area Network ( 

→ LAN)/Wireless Local Area Network ( → WLAN) links as well as

convenient protocols like IP, Internet Control Message Protocol ( → ICMP), Simple Network Management Protocol ( → SNMP) The program support,

configuration and tuning of the intelligent network is realised with the use ofmobile devices (smartphones and tablets), mobile applications and throughoffered web services running in a cloud environment This leads to a smartenvironment in which all company device capabilities are used in

combination to their full extent to ensure autarky with high security and

privacy, but still on-demand scalability beyond the company’s realm, andhigh energy efficiency with inclusion of all local energy sources and jointbrokering of power and computing supplies We name the outcome of thisscenario a smart grid environment

Fig 1.4 Hybrid cloud/heat providers: combination of smart grid, clouds and Heating, Ventilating and

Air Conditioning ( → HVAC)

Scenario 2 Energy recycling in data centers Due to use of today’s

powerful high-end servers within the contemporary data centers with theinstalled broadband optical links (e.g Fibre Channel), a significant amount ofheat stands out as a harmful by-product Some companies occupy themselves

already with the mentioned problem and are developing their own solutionsfor the disposal of heat excesses for domestic heating and air-conditioning

facilities, the so-called HVAC Among them are hybrid cloud and heat

product providers [8] These companies have a portfolio of several

corresponding products and solutions (Fig. 1.4), inter alia there are cloudinfrastructure and platform services, and heat products, representing an ownsmart grid with inter-connected services The clients use the in-door locatedservices of virtual computing centers, standardised cloud services like

Infrastructure-as-a-Service ( → IaaS), Software-as-a-Service ( → SaaS) and Platform-as-a-Service ( → PaaS) Among them, there are popular

applications causing a significant amount of heat, from computing servicespowered by cloud stacks, virtualised operating systems, and add-on serviceslike databases and cron jobs Redundant heat as a “by-product of processing”

is withdrawn via servers in 19 ′ ′ -racks in the energy storage, which providescirculation of hot water in the pipes within a building and heating of potable

water The central system for HVAC facilities is supported via use of Power over Ethernet ( → PoE), as well as wired and wireless automation local-area and piconets like LON, KNX, ZigBee, EnOcean The mentioned technical solution provides a lower PUE value down to 1.05 or, correspondingly, an efficiency 1/PUE up to 95 %, compared with the conventional grid/cloud-

solutions, where it is necessary to remove the excess heat as by-product, toinstall more air-conditioning devices and provide them with power supply.Similarly, a growing number of data centres world-wide are inter-

connected with municipal utility providers to funnel their excess heat intopipes which lead to central heating systems of housing areas Interestinginstallations exist in Helsinki, Finland, where servers located beneath theUspenski cathedral in the Academica/Telecity Group server centre heat 500homes as by-product More servers located in a shielded building insideanother building, a former electricity station now hosting the Suvilahti datacentre, even offer heat and warm water for 4500 households

Scenario 3 Low-cost and energy-efficient on-board microcontrollers for pico-services But none of the above-mentioned computing systems is

energy-efficient enough to meet the ambitious goals set by environmentalistsand to some degree even political agendas Switzerland, for instance, is

committed to reduce the emissions in 2030 to just 50 % of those in 1990.Germany intends to reduce emissions until 2020 to 60 % The electricityconsumption in data centres is in the MWh area and even for tiny

computations, a power-hungry large machinery of hardware and supportprocesses is needed Energy-efficient solutions can be provided via small,low-cost and low-energy on-board processors on which pico-services, such aslambda services, are executed on demand The electricity consumption getsreduced to the kWh area or even less Low-energy home intelligent nodes (3–

10 W) for private cloud solutions, file servers, web servers, multimedia homecentres etc can be placed on the low-cost energy-efficient on-board

microcontrollers like Arduino, Raspberry Pi or Intel Edison as a trade-offsolution They offer a cheap alternative and symbolise a step-by-step shift to

the IoT But in order to maximise their potential, an appropriate service and

application platform will be needed

An appropriate solution will be the Raspberry Pi

on-board-microcontroller (first deployed in 2011 in Cambridge, UK) with only creditcard dimensions, in a pod like a matchbox and with the following

characteristics [4]: A 700 MHz processor, a modest amount of main memory

up to 1 GB, external storage on an SD card, an Ethernet connection or a

wireless link through a USB dongle, and around 3.5–5 W power

Multimedia environment: XBMC Media Center;

Multiple audio and video formats (codecs) as well as low power

The newest Raspberry Pi 2 Model B acts as a mini-PC with 6 times theCPU performance due to a tact frequency of 900 MHz and a quad-core

architecture being oriented to the Windows Developer Program for IoT But

even more energy-efficient boards are upcoming, including the Genuino withthe Intel Curie chip and the Pine A64 which even runs on a 3.7 V Lithiumbattery

How to read on This was a quick chapter! The next ones will have more

depth, as they convey the actual knowledge about the mentioned areas In thesecond chapter, the development of network systems will be summarised andpresented with historical and contemporary systems In the third chapter,

clusters and parallel computing will be focused on Virtualised systems andclouds will follow in the fourth chapter Chapter number five will step intothe physical world and contains information about smart grids, smart thingsand smart fog While the sixth chapter will present mobile communicationtrends, the final seventh chapter talks about security aspects in a broad

meaning With such a spectrum of topics, the reader should then be able tounderstand both old and new large-scale systems

References

1 Bence Bakondi, Péter Burcsi, Péter Györgyi, Dávid Herskovics, Péter Ligeti, László Mérai,

Dániel A Nagy, and Viktória Villányi A P2P Based Storage System with Reputation Points and

Simulation Results In Central European Conference on Cryptology (CECC), Budapest, Hungary,

May 2014.

2 Markus Bloesch netrics übernimmt Umweltverantwortung: Cloud Computing und Hosting aus Datacenter mit Ökostrom aus dem Wasserkraftwerk Hagneck online: https://www.netrics.ch/2015/ 12/03/cloud-computing-hosting-mit-oekostrom/ , 2015.

3 Alisa Davis Equinix Goes 100 % Renewable with 225-MW Wind Energy Purchase online: http:// apps3.eere.energy.gov/greenpower/news/news˙template.shtml?id=2082 , 2015.

4 Raspberry Pi Foundation Raspberry Pi Hardware online: https://www.raspberrypi.org/

documentation/hardware/raspberrypi/README.md , 2015.

5 Thomas Loruenser, Charles Bastos Rodriguez, Denise Demirel, Simone Fischer-Huebner, Thomas Gross, Thomas Langer, Mathieu des Noes, Henrich C Poehls, Boris Rozenberg, and Daniel

Slamanig Towards a New Paradigm for Privacy and Security in Cloud Services 2015.

6 Eric O’Shaughnessy, Jenny Heeter, Chang Liu, and Erin Nobler Status and Trends in the U.S Voluntary Green Power Market Technical Report NREL/TP-6A20-65252, National Renewable Energy Laboratory, 2015.

7 Eric Schmidt The Internet Will Disappear World Economic Forum, via CNBC TechBet, online video: https://www.youtube.com/watch?v=Tf49T45GNd0 , 2015.

8 Rene Marcel Schretzmann, Jens Struckmeier, and Christof Fetzer Cloud&Heat Technologies online: https://www.cloudandheat.com/ , 2011/2014.

9 Internet Society RFC Editor online: http://www.rfc-editor.org/ , 1998.

10 Yevgeniy Sverdlik and Karen Riccio Special Report: The World’s Largest Data Centers online:

http://www.datacenterknowledge.com/special-report-the-worlds-largest-data-centers/ , 2010.

(2)

© Springer Fachmedien Wiesbaden GmbH 2017

Andriy Luntovskyy and Josef Spillner, Architectural Transformations in Network Services and

Distributed Systems, DOI 10.1007/978-3-658-14842-3_2

2 Architectural Transformations in

Distributed Systems

Andriy Luntovskyy1 and Josef Spillner2

Berufsakademie Sachsen, Dresden, Germany

Service Prototyping Lab, Zürcher Hochschule für Angewandte

Wissenschaften, Winterthur, Switzerland

Andriy Luntovskyy (Corresponding author)

Email: andriy.luntovskyy@ba-dresden.de

Josef Spillner

Email: josef.spillner@zhaw.ch

Keywords Client-Server ( → C-S) – Peer-to-Peer ( → P2P) – Central

Database ( → CDB) vs Distributed Database ( → DDB) – Transactions

The timeline given in the first chapter embodies the perspective of humansusing and benefiting from services In this chapter, we now dive under thehood of this development and take a look at the service software

implementations with a special focus on basic principles of complex

distributed services which fulfil the requirements for modern cloud and fogapplications Over the last two decades, we have been able to observe

significant architectural changes in distributed systems and networking

applications which will be reflected in the text There are also mostly

orthogonal shifts towards higher reliability, efficiency, scalability and

information security, as well as other benefitial non-functional characteristics

The chapter covers general software and system architectures, discussescluster and cloud systems as well as peer-to-peer topologies, along with

concrete system examples, and highlights the topics of performance

optimisation and transactions as well as distributed databases

2.1 Software Architectures and Communication

Patterns

Among the most well-known conventional service architectures for softwareapplications are the client-server model and the n-tier model In the client-server model, a client connects to a server to exchange messages with it inorder to achieve a certain goal In the n-tier model, multiple client-serverconnections exist in a chain Let us consider an integrated example

Example 2.1.

A distributed software application for e-commerce has frequently a rathercomplex hierarchical structure, called n-tier, which is created with the aim ofperformance optimisation, and includes programmatic interfaces linked withnetwork protocols An example of a system for e-commerce is depicted inFig. 2.1 The application 1 for a purchaser (client) interacts with the virtualshop, i.e application 2 (online shop) via a web server with the attached

application server, which provides data preprocessing for purchase orders.The application server for the purchase order preprocessing is connected tothe next two application servers One of them is aimed at store managementwith maintainance of store tables, the other one at administration of customerdata The application 3 supports the communication of the online shop withthe suppliers via a dedicated communication channel which is connected to

an application server as well as the supplier database Communication

between the applications 2 and 3, i.e online shop-to-suppliers, is performedwith use of a corresponding channel provided by the platform Thus, we seethe advancement of typical application architectures to distributed systemswith client-server and n-tier architectures [5, 7, 8]

Fig 2.1 Example system: e-commerce (Source: [5 ])

As it was shown in [5, 8], multi-tier architectures nowadays are widely

deployed in distributed applications:

3-tier: structure is more complex, leading to higher scalability, preferredfor complex applications;

2-tier: two-tier structure (user interface and host), is simpler but lessflexible (Fig. 2.2)