APPENDIX – POSTERS AVAILABLE VIA http://wiki.n4c.eu/wiki/index.php/Posters_Index

Contribution by

Luleå tekniska universitet

Country

Sweden

Posters

• • •

Complexities of the access to the internet – examples from Sweden and Uganda LTU's DTN Tests Facts in Short DTN PODCAST

(Coordinator)

Albentia Systems, S.A.

Spain

•

Enhanced WiMAX on N4C

Universidad Politécnica de Madrid

Spain

•

Animal tracking application with self-powered secondary nodes

Intel Performance Learning Solutions Ltd.

Ireland

•

Intel IPLS – presentation of the Atom based village router

Northern Research Institute Tromsö AS ITTI Ltd.

Norway

•

Pervasive DTN Applications

Poland

•

Simulation platform with use examples

Instituto Pedro Nunes

Portugal

•

Integration Challenges and Deployment Strategy

MEIS storitve za okolje d.o.o.

Slovenia

•

Slovene tests 2010

Tannak AB

Sweden

•

Animal tracking application with battery powered Secondary, Primary and Transmission nodes

Power Lake AB

Sweden

• •

N4C Business and Deployment Planning Public Dissemination

• •

PRoPHET DTN Routing DTN Fundamentals

Folly Consulting Ltd.

United Kingdom

Networking for Communications Challenged Communities: Architecture, Test Beds and Innovative Alliances Contract no: 223994

Complexities of Access to the Internet. Examples from Sweden & Uganda Majority of the world’s population does not have access to the internet. Several initiatives underway to address this situation Is there a best approach to finding solutions for the billions with no access? And what is access? Imagine different layers affecting access channelling into each other to create moments of what may become access or non-access. Not all the layers may apply to a given access situation, sometimes the presence of two layers is adequate and other times the absence of just one may be what impedes access. Looking at the graph, there is one central intersection but there are several other intersections which add to the reality of access. Caroline Wamala

Networking for Communications Challenged Communities: Architecture, Test Beds and Innovative Alliances Contract no: 223994

• • •

The presence of the technology does not translate into automatic access The access graph shows the concept as an intersection of the social and the technical For example: During field research in Uganda Caroline Wamala met Nora who lives in a rural area. Nora says that she does not know how to use the computer or the Internet. But Nora regularly uses the Internet to communicate with her son who is overseas, “I ask others to do it for me”. Nora’s indirect access requires an additional human node that enables what could have been non-access for her. A similar example is provided by Fritz-Åke Kuoljok field test manager of N4C in Sweden, who informed of a scenario where a father relied on his child to help use the internet. Even though the father himself can not use the technology his access is not impeded by this limitation, but he acquires indirect access through his child. Caroline Wamala

Networking for Communications Challenged Communities: Architecture, Test Beds and Innovative Alliances Contract no: 223994

•

On the other hand Caroline experienced moments of non-access when conducting the field work in Uganda. Her user characteristics include having the financial ability and technical know-how (see access graph).

•

Yet, her ability to use the technology could not help when the infrastructure did not work. Caroline is regarded as having access, where Nora is not, and yet the situation for either is not static but involves perpetual negotiations with different layers of access.

•

The N4C project, in addressing non-access for the rural remote areas in Europe in its approach considers the social (community life-style) jointly with technical solutions in developing a working-solution for communications challenged areas.

•

A socio-technical approach (Cummins 1978) to providing internet access where there is none, takes into account access in its complexity; what does the community in question want how are they organised, etc. simply providing ‘technical’ access does not guarantee use. Caroline Wamala

LTU's DTN Tests Facts in Short DTN-IT services provided on the field: - Email (sending/receiving) - SMS (sending out from the test field) - Not-So-Instant-Messaging (used for communication within the test field) - Podcast (access to daily audio news from different radio stations) - Web-caching (web access to daily snapshots of favorite web pages) - Weather monitoring (access to the data from a weather station located on the field) - Online DTN system monitoring (automatize collection of all the daily sent log files from the field, including web-cam images on the Internet)

The last summer test in numbers: -number of DTN nodes: 18 -test period: 53 days -number of users: more than 60 (20 registered users, 40 guest) -supporting for platforms: Linux, Embedded Linux, MS Windows and Apple OSX -number of sent out bundles: 6107 bundles -bundle delivery rate: 89% -average bundle delivery delay: 2 days -users sent out: 1359 Emails, 328 NSIMs, 167 SMS

The main outcome from LTU's DTN tests: -collection of valuable real-world connectivity and network traffic traces that can be used in further research -feedback from involved users collected through interviews -proved usability of the DTN architecture and the DTN services in terrestrial scenarios -improved dynamic ProPHET routing protocol -more reliable hardware design of the outdoor DTN station

Author: Samo Grasic, Luleå University Of Technology 2011

LTU's DTN Test Topology Helicopter base in Ritsem B1 C

A

LTU

B2 Daily scheduled helicopter flights (Ritsem-Staloluokta-Ritsem) Distance: 60km

E

WLAN link Distance: 1km

H J

G Charging station (for test participants): 2x 12V 100Ah battery 3x 60W Solar panels

I

Staloluokta (tourist cabin)

N K

M

Staloluokta (Sami village)

L

O

D

Helicopter base in Stora Sjöfallet

N4C-LTU, the N4C test team from Luleå University of Technology, has developed and field-tested a “store and forward” ubiquitous application that provides DTN end-users with daily audio shows such as news and weather.

DTN PODCAST APPLICATION N4C-LTU Podcasting provides a broad selection of global digital audio broadcasts for the DTN network. The arctic Sámi camp Staloluokta is a place where no electrical power infrastructure or cell coverage exist. This DTN application assumes that endusers have computers that are charged with solar & wind power generators.

The end-user web GUI

The nomadic end-users can regularly receive podcasts with the help of DTN data mules.

DTN PODCAST APPLICATION How DTN podcasting works Our N4C-LTU gateway supports the subscription of podcasts. It sends out the podcast files to the DTN daily. Whenever DTN nodes connects they will transfer the files. End-users will access them trough N4C-LTU GUI and listen to them any time. The N4C-LTU podcasting also allows the end-user to request podcasts from a list (while in the Communications Challenged Region) for adding to the gateway subscriptions.

Enhanced WiMAX on N4C Interoperable WiMAX Repeater • Portable solution for long distance links as part of DTN networks • > 300km with a repeater using 60cm antennas relaying on DTN to increase availability • Field tests performed in Spain in real application scenario with real users (> 6 months) • Doubles the distance of the link or increases data capacity • Enhanced coverage area • Allows connection in NLOS scenarios

N4C’s Repeater Lab Tested

Real Field Tests in Spain

Enhanced WiMAX on N4C Mesh Node For Static Long Distance Wireless Sensor Networks • Base and Subscriber Station modes integrated in a single device • “On the go mode-switch” allows connectivity between all nodes • XML-RPC API provided for network organizers • Algorithm for real network implementations developed

N4C’s Targeted Mesh Application

WiMAX Link Distance Enhancement Using Multiple Antennas

N4C’s MISO prototype

• MRC Algorithm developed for WiMAX applications • Adaptation of MISO technique to operate with real WiMAX products • First prototypes for lab testing ready

Networking for Communications Challenged Communities: Architecture, Test Beds and Innovative Alliances Grant Agreement: 223994

Animal tracking application with self-powered secondary nodes •The network considers three types of nodes of different functionalities: • Secondary units / nodes: simple devices with kinetic generators (no batteries are required) to transmit their identification • Primary (leader) units / nodes: equipped with GPS and batteries to collect the transmissions of the secondary units and behave as ‘data mules’ until delivered to the hot spot • Hot spot: gateway with the rest of the network. It will be also equipped with long range links as GSM / WiMAX 14th of April 2011

www.n4c.eu

1

Networking for Communications Challenged Communities: Architecture, Test Beds and Innovative Alliances Grant Agreement: 223994

Main features •Secondary node •Motivation: Decrease the number of batteries in the system. Replace the standard approach with batteries by a kinetic generator that makes use of the movement of the animal. •Most of the animals are equipped with these devices. • Random transmissions • Radio Range about 100 m •Very simple radio transmitter in the 433 MHz band using an FSK modulation

The kinetic generator is a plastic tube with a moving magnet inside and a wire coiled around: the converted power (Lenz’s law) when the magnet moves is used to energize and trigger the transmitter 14th of April 2011

www.n4c.eu

2

Networking for Communications Challenged Communities: Architecture, Test Beds and Innovative Alliances Grant Agreement: 223994

The secondary node

Kinetic Generator

Radio Transmitter 433 MHz, FSK Transmits just the Identification number

14th of April 2011

www.n4c.eu

3

Networking for Communications Challenged Communities: Architecture, Test Beds and Innovative Alliances Grant Agreement: 223994

Main features •Primary (leader) nodes • Just few animals are equipped with these devices. • More complex including a GPS and batteries • Collect (433 MHz receiver) the transmissions from the secondary nodes assigning a time stamp and their own GPS location • They act as ‘data mules’ until they come close to the hot spot (opportunistic transmission) or just periodic downloading • Equipped with a second radio transmitter in the 166 MHz band using also FSK modulation to increase the coverage to the hot spot

GPS

Radio Receiver 433 MHz, FSK Radio transmitter 166 MHz, FSK

14th of April 2011

www.n4c.eu

4

Intel IPLS Mainly working on WP4 and WP5 , Intel worked with TCD on the development of the DTN village router, this included the working on the solar power and power management of the router and documenting the build. They also assisted in WP6 and completed WiMAX testing in Ireland(www.cordis.lu/fp7).

As seen from the DTN topology, the village router is at the core of any DTN. This router needs to be stable and have robustness and be generally accessible to all modes of transporting data.

Text: More text here if needed

The DTN router collects the requests made by the DTN nodes and then sends the requests to the data mule, that takes requests to the DTN gateway.

Intel IPLS The Router was made up of Intel Atom Based SBC from a company called Eurotech. We utilise sensors to manage the power and uptime of the system. We had ‘managed uptime’ up-to 94% during the summer 2010 trial. We utilised the systems in Ireland for winter trails also. Alls systems survived .

The village routers were placed in three locations. Staloluokta, Skuolla, Sallohaure and Vastenjaure. These were chosen because of potential high activity and safety. They all linked back into TCD data mules and VPN from Ritsem and Sjofjallet.

Text: More text here if needed

Pervasive DTN Applications

The applications are in contact with the N4C DTN Hotspots (R2 or LTU), and the helicopter (m3 – data mule) brings the data to/from the Internet once daily.

Hiker’s App Hiker’s App shows the connected devices (reachable/unreachable), and their locations (Distance/Bearing/Speed/Latitude/Longitude).

N4C Hiker’s App has a Photoblog which syncs photos taken, and sends them via DTN to the N4C web server.

The photos are synced automatically when other devices are reachable. The photos can also be manually synced when in contact with the N4C DTN Hotspots or other devices (Nokia N810, Nokia N900 or Linux Netbooks). Hiker’s App Photoblog

Software Architecture N4C has developed and field-tested “store and forward” ubiquitous applications that provides DTN end-users with services such as Email, Web-services, Hiker’s, Herder’s and Meteo Apps. The Software architecture is divided in four different tiers: 1. The Application layer called Hiker’s PDA contains the applications. 2. The middleware layer. 3. The DTN layer which may provide communication to the Internet or to different CCRs. 4. The link layer which is managing the different communications (e.g. WiFi, WiMax and Bluetooth).

N4C Software Architecture The implementation of the Hiker’s App is based on Python. The hybrid synchronization service in the middleware layer uses XML-RPC for synchronization of meta-data, and http for synchronization of user data. The auto-discovery function uses IPv6 and multicast.

Simulation  platform   The simulation platform is part of the integration platform of Work Package 7, addressing the needs of a professional developer. Objective: ‡ support integration testing of networking stack and application software Features: ‡ mix real and virtual network nodes of different types (PCs, netbooks ± Linux, routers ± OpenWrt, smartphones ± Maemo, Android, Symbian) ‡ set up, replicate, and modify various networking scenarios easily ‡ visualize node movement and contacts, and trace network traffic ‡ interact with the nodes and the network during simulation

Another network simulator? ‡ a tool for system integration rather than theoretical research ‡ emphasis on facilitating observation of actual software and hardware being integrated 1/4  

Platform  internals   Based on open-source software components, including: ‡ ns-3 network simulator: extensible simulation engine ‡ VirtualBox virtual machine system: host to virtual nodes with full applications ‡ Linux containers resource virtualization system: host to transport-only virtual nodes ‡ OpenVPN server: for remote nodes over the Internet to participate in simulations ‡

‡ ‡

‡

Adds C++, Python, and shell code for scenario models, networking setup, system configuration, automation, and GUI Connects virtual as well as real nodes Networks with delays and disruptions modelled in ns-3 up to layer 2 (data link), including topology and mobility, data transfer ranges, rates, and error characteristics Higher network stack layers (incl. routing) not simulated ± run by actual software being tested 2/4  

Use  examples   Right: testing basic DTN2 applications in 3 virtual machines and 1 real N900 phone Bottom left: testing N4C +LNHU¶V PDA application in the same node set Bottom right: visualization of testing a DTN network with a mix of 93 static and mobile nodes

3/4  

Benefits   Software and hardware being integrated can be tested in large, complex DTN networks: ‡ ‡ ‡ ‡

networks are simulated, easily modified, scaled, and replicated software is real hardware can be real or virtual simulations can be interacted with as well as analysed off-line

Applications can be tested in networks of hundreds of nodes simulated on a PC instead of engaging huge hardware resources, manpower, and time for each test. The code is open-sourced and available at http://info.n4c.eu/code/simulation-tools Contact: {Lukasz.Kiedrowski,Krzysztof.Romanowski}@itti.com.pl

4/4  

Integration Challenges HTMLrequester

Pymail

Prophet

Meteorological service

DTN

DTN

DTN

DTN

Several modules, or services, were developed within N4C project but these modules were not made sharing a common platform. The challenge of the system integration is to develop a form to all the modules to work together in a common ground. This will allow all kind off users, from the less to the most experienced, to be able to use all the services provided by N4C. To do this two very important things had to be made: • Deployment Procedure • Integration Platform The deployment consists of a simple way of installing all the modules needed and the integration platform is the common ground developed to shelter all the modules.

Deployment Strategy

The aim is to deploy and disseminate N4C technology to the communities where it is not economically viable to provide internet (and its inherent online resources) email access to population in remote areas Every deployment strategy has three very important stages: Installation, Configuration and Usage. - Installation: Installation packages to several platforms were developed to make all the installation procedure simple. All the installations can be done by done Graphical User Interface (GUI) - Configuration: All the configurations needed for the end user can be done through the Integration Platform - Usage: The usage is very intuitive making the platform suited for all kind of users. Help is provided about all the problems that the user may find.

The Integration Process PRoPHET

HTMLrequester

Meteorological

Pymail

...

Integration Platform The Integration Platforms roles are: - Gather all the modules in a single platform; - Be off simple usage; - Be easily expandable to new modules; - Control DTN on the user side; - Be accessible to developers - Cross-platform - Do the interaction with Outstations

A high level of interoperability has been achieved with a Python based Integration Platform that can be used in every Linux based OS. Being Python based it allows new modules to be easily integrated without conflicts

Screens

A explorer was embebed into the Integration Platform making it possible, for example, to use HTMLrequester through it or to access help pages on the Outstation

Slovene tests Summer tests 2010 Slovenia

1/4

Slovene tests Main topology

μ-GaRaMo topology

2/4

Slovene tests Devices used

Meteorological station with a DTN router

μ-GaRaMo radiological station with SymbioNode node USB-key connectivity

Meteorological station with a DTN router – exterior part

Webcam picture (wildlife observation)

3/4

Data mule in car

Meteorological station “Pustice”

Slovene tests Traffic evaluation

4/4

Networking for Communications Challenged Communities: Architecture and Benefits Grant Agreement: 223994

Animal tracking application with battery powered Secondary, Primary and Transmission nodes Tannak AB Karin Kuoljok, Susanne Spik www.tannak.se

2011-04-10

www.n4c.eu

1

Networking for Communications Challenged Communities: Architecture and Benefits Grant Agreement: 223994

Summary The Tannak system consists of three main parts; a simpler secondary equipment for the majority of the animals, a more sophisticated primary equipment for collecting information from the majority, attached to fewer animals, and base stations that collect and distribute information from the primary units to the user’s computer. The Tannak system is to some extent delay tolerant, as the primary units and base stations store information until contact is established with the receiving part in the aggregating chain of communication. The herd or group of animals is positioned by using GPS in the primary units. Data from each individual animal consists of identity, movement indicator, battery status and ambient temperature. The user can monitor the individual status of each animal, the general environment (temp) and the position of the herd; if the herd is within defined bounds or close to off limits. The movement indicator shows animals not moving (dead or sick) or lost collars. Presenting the herd information remote at home will reduce the need for searching and physically monitoring the herd by using snowmobile, helicopter or motorcycle, saving costs and environment. The ID function enables the user to individually trace an animal during its lifecycle, which is important for the food industry.

434 MHz

Secondary Unit, encapsulated

2011-04-10

150 MHz

Primary Unit, with GPS window

USB

HotSpot Unit, with GPS on top

www.n4c.eu

Computer, Connected to HS, Directly or over Internet

2

Networking for Communications Challenged Communities: Architecture and Benefits Grant Agreement: 223994

System Overview –Secondary Unit (SU) transmits individual information to Primary Unit (PU) –Primary Unit collects and transmits information including GPS position to a HotSpot (HS) –HotSpots synchronize information between HotSpots within range –HotSpot transmits data over USB to a modem connected to the Internet, or to a Computer –A Computer (PC) collects, stores and displays information

SU

SU

SU

SU

SU

SU

HS Modem

PU HS

SU SU SU SU SU SU

2011-04-10

PU PC

HS

www.n4c.eu

3

Networking for Communications Challenged Communities: Architecture and Benefits Grant Agreement: 223994

Secondary Node The Secondary Unit (SU) has relatively simple electronics and is enclosed in a IP65 encapsulation The SU transmits the following information to the Primary Unit (PU) on predefined intervals: –Identity (RFID), Temperature, Movement indicator, Battery status It uses 434 MHz radio. The range is approximately 200 meters. The transmission intervals can be set from 10 minutes up to 8 days

Secondary unit electronics, shown with and without the battery, and one waterproof encapsulated unit, including antenna

2011-04-10

www.n4c.eu

4

Networking for Communications Challenged Communities: Architecture and Benefits Grant Agreement: 223994

Primary Node The Primary Unit (PU) and the HotSpot (HS) has common electronics, but slightly different software and will be enclosed in a IP65 encapsulation The PU communicates with the SU over the 434 MHz radio and with the HS over a 150 MHz radio The range to a HS can be up to 6 km. The PU transmits the following information to the HotSpot (HS) on predefined intervals: –Collected SU information, Temperature (mean), Battery status and GPS position The transmission intervals can be set from 10 minutes up to 8 days

Primary Unit / HotSpot Unit board is shown without its screen plate and encapsuling The board is 47x67 mm. Each square on the backing paper is 5 mm.

2011-04-10

www.n4c.eu

5

Networking for Communications Challenged Communities: Architecture and Benefits Grant Agreement: 223994

Transmission Node, HotSpot The HotSpot (HS) and the Primary Unit (PU) has common electronics, but slightly different software and will be enclosed in a IP65 encapsulation The HS battery can be charged via the USB port, which also is used for communication to a modem or a computer. The HS communicates with the PU over a 150 MHz radio, range ca 5 km The HS synchronizes collected information and own position with other HS:s within range over the 150 MHz radio and the range can be up to 10 km, depending on topography. Any HS within range can be used for loading collected information into a computer or a mobile HS

Primary Unit / HotSpot Unit board is shown without its screen plate and encapsuling The board is 47x67 mm. Each square on the backing paper is 5 mm.

2011-04-10

www.n4c.eu

6

Networking for Communications Challenged Communities: Architecture and Benefits Grant Agreement: 223994

Additional The Tannak system allows the user to access collected data in different ways; via a mobile HS connected to a computer, via a fixed HS connected to a computer or via Internet over a modem connected to a HS. A chain of HS:s can propagate information to a suitable access or information delivery point. The user is able to monitor the movements of the animals from a remote location, thus saving time, cost of gas for snowmobile travelling or helicopter flying in search of the herd. Individual users can either have their proprietary system or share HS:s within a community for better coverage. Below a tracking example from the winter tests Feb 2011

2011-04-10

www.n4c.eu

7

N4C BUSINESS AND DEPLOYMENT PLANNING

My research aims to describe, test and evaluate business models suitable for N4Cs business planning and to identify suitable models for governance of N4C outcome. Research is rooted on three pillars:

“Business” means all form of industrial and commercial profit-seeking activity But there is hardly any business to develop connections to the internet in some rural areas! N4C business planning have to take that into account and develop models suitable for communication challenged areas.

1.Business models; both generic models for ICT devices and services and business models for Open Source Software (OSS). For OSS I have tested models developed by Hecker. 2.Governance models that I have tested for N4C outcome (the DTN based internet access, test beds and applications) is based on recommendations by professor and Nobel gloriat Ellinor Ostroms´ models for managing Common Pool Resources (CPR). My research question is: what can we learn from those governing models when we develop models for N4C outcomes? Test have been conducted on usage of an economic association for the DTN based internet access. 3. Research methods are user-driven interactive research with participatory design methods.

1/12

How N4C DTN architecture works 1. A mast with long reach is placed in the area. 2. Person1 is going on the snowmobile and is within reach of the mast so he can surf and mail on a wireless computer (PDA, mobile etc.) 3. Not yet visible for Person1 the DTNsoftware in his/her computer is downloading mail to Person2 and Person3. 4. When Person1 is passing by a Wireless Local Area Network (WLAN) station driven with solar cells in the terrain Person2´s and Person3´s mail are copied to the station.

Illustration: Bo Reinerdahl

5. When Person2 is passing the WLAN station the DTN is copying the hidden mails to Person2‟s unit. The same way mails are spread to other WLAN stations. Mails that are finally delivered are erased gradually. 6. The WLAN-stations are also updated when regular flights or helicopters with DTN equipment is passing. In that way both in- and outgoing mail can be monitored many times per day. 2/12

Governance of DTN based internet access with usage of Ostroms CPR models - tests on the DTN based internet access

“Providers” are European Commission , LTU, TCD, UPM and central government in participating countries.

“Producers are those who actually construct, repair or ensure the long-term sustainability of the DTN based internet access. In N4C it includes LTU, TCD and UPM local SMEs, NGOs and private persons.

“Appropriators” are citizens in Gällivare and Jokkmokk, SMEs (tourist companies, helicopter transport enterprises etc), NGOs (Sámi organisations and their cultural associations) and those who uses DTN based internet access for communication. Appropriators can also be local authorities.

All three; Appropriators, Providers and Producers can be the same individuals. Source: Ostrom, E (1990), Governing the Commons: The Evolution of Institutions for Collective Action. New York: Cambridge University Press.

3/12

N4C business modelling Questions to be asked: (Markides 1999) for N4C outcome: 1. Who is the target customer? 2. What products or services is N4C offering? 3. How can these services be delivered in the best way to the customer? N4C OUTCOME •DTN based Internet access in Swedish Lapland and in Slovenia) • Test Beds (in Swedish Lapland and in Slovenia) • MEIS´Applications (3) 1. Meteorological station - AMS-DTN for wind power 2. Meteorological station with extended autonomy for critical situations measurements AMS-cs of possible air pollution dispersion into the atmosphere 3. Symbinode gamma dose rate measuring device - u-garamo Norut´s Application: Apps for Hikers PDA a portable handheld computer device for Communication Challenged Regions Tannak´s Application: Animal Tracking UPM / Albentia´s: Wireless technology for sensitive areas LTU / TCD /IPLS/ Folly´s: Improvement of usability of DTN based internet access (Routing, podcast, NSIM, web cache, e-mail, hardwares) 4/12

Different stages of N4C outcome Based on an idea by Pacific Coast Research Inc. presented in Schlueter Langdon, C and Hars, A In Open Source Software Business Models and Customer Involvement Economics. M Business Scenarios O D Descriptive analysis E L

Quantitative

Rollout

Conceptual model

Estimates

Decision

Test Bed

N 4 C O U T C O M E

Qualitative Assessment

MEIS – Application Symbinode gamma dose rate measuring device - u-garamo

Test Bed

Swedish Lapland NORUT: Apps for hikers PDA LTU / TCD /IPLS/ Folly: Improvement of usability of DTN based internet access , i.e. Routing, podcast, NSIM, web cache, e-mail, hardwares

MEIS – Application (Meteorological station - AMSDTN)

UPM / Albentia Wireless technology for sensitive areas

Slovenia MEIS – Application (Meteorological station with extended autonomy for critical situations measurements AMS-cs)

Tannak Animal Tracking

MEIS Test Bed is in rollout stage whereas Test Bed Swedish Lapland is in Qualitative Assessment Phase, Norut´s apps are in Qualitative Assessment Phase whereas MEIS AMS-cs is in rollout phase.

5/12

Test of business and governing model for DTN Based Internet Access in Swedish Lapland

Based on: Osterwalder, A, Pigneur, Y, Tucci, C, (2005) and Ostrom, (1990) Pillar

Business Model/ Building Block

Product

Value Proposition Customer Target Interface Customer /User Distribution Channel Relationship

Infrastructure Management

Financial Aspects

Description

Access to the internet with DTN based internet access including e-mail and web browsing service on a basic level. Appropriators: people living part of the year in Swedish Lapland and/or working or visiting the area. Producers: citizens living in the area, researchers (LTU, etc. ) universities and incl. Gällivare and Jokkmokk dev. agencies. Services on the DTN based internet access will be distributed by Producers gathered in an Economic Association for N4C Delay Tolerant Network in Swedish Laponia. The economic association will be open for anyone but it will be expected to run by Providers and Producers. Providers are funders or donors (ICT industry). Producers are citizens in the area but also researchers and similar staff at the LTU, TCD, UPM including the two municipal development agencies in Gällivare and Jokkmokk.

Value Configuration Core Competency

Demands limited budget (in cash) as the N4C DTN based Internet access use Commercial-Off-The-Shelf (COTS) and Open Source Software (OSS) (Apache 2.0). The Producers competencies (citizens in the area but also researchers and similar staff at the LTU, TCD, UPM including the two municipal development agencies in Gällivare and Jokkmokk). The Producers role is to maintaining the infrastructure and run the Economic Association. Partner Providers and Producers will be member of the economic association operating the Network DTN based internet access. Cost Structure Monetary consequences of the DTN based internet access will be costs for updating and maintenance of the equipment and work on improving the system. Cost for extra test equipment etc. will be carried by the clients who need that extra infrastructure. Revenue The business model service support seller (Hecker´s) will be used for the DTN Model based internet access. Revenues for updating and maintenance will be gained from other sources than from licence fees, i.e. membership fees and donations from ICT industry etc. Other sources of revenues will be from usage of the DTN internet based access on the test bed and as infrastructure for the applications (both those available now and developed in future).

6/12

Test of business and governing model for DTN Based Internet Access in Slovenia

Based on: Osterwalder, A, Pigneur, Y, Tucci, C, (2005) and Ostrom, (1990) Pillar

Product

Business Model/ Building Block Value Proposition

Customer Target Interface Customer/ User Distribution Channel Relationship Infrastructure Management

Financial Aspects

Value Configuration

Description

Access to the internet with DTN based internet access including e-mail and web browsing service. Appropriators are users living in Slovenian mountain who will benefit. Provider is MEIS and their collaborating partners. Producers are citizens living in the area but also researchers and similar staff at MEIS who are prepared to work on the DTN based internet access . Services on the DTN based internet access will be distributed by MEIS and their partners. The organizations taking part in running the DTN based internet access are MEIS and their partners. The N4C DTN based Internet access demands limited budget (cash) as is uses Commercial-Off-The-Shelf (COTS) hardware‟s and Open Source Software (OSS) (Apache version 2.0 licence).

Core Competency

MEIS is the “carrier” of the DTN based internet access. Their competencies are in ICT and energy.

Partner Network

Appropriators: people living in the area; Provider: MEIS, Producers: MEIS, transport companies, etc.

Cost Structure

Monetary consequences of the DTN based internet access will be costs for updating and maintenance of the equipment and work on improvement of the system administrated by MEIS. Cost for extra test equipment etc. will be carried by the clients who need that extra infrastructure. A business model called service support seller (Hecker) will be used by MEIS and their partners. Revenues for updating and maintenance will come from other sources than licences. Main sources of revenue will be usage of the DTN internet based access

Revenue Model

on the test bed and as infrastructure for MEIS own applications.

7/12

Test of business and governing model for Test Bed Swedish Lapland Based on: Osterwalder, A, Pigneur, Y, Tucci, C, (2005) and Ostrom, (1990)

Pillar

Product

Business Model/ Building Block Value Proposition

Customer Target Interface Customer /User Distribution Channel Relationship

Infrastruc Value ture Configuration Managem Core ent Competency Partner Network Financial Cost Aspects Structure Revenue Model

Description

The Test Bed will offer services in Communication Challenged Communities with test area where the climate is harsh and with testing performed on a DTN based internet access as infrastructure. Universities, SMEs, ICT industry and authorities. Test Bed services can be distributed by an Economic Association in Swedish Lapland run by SMEs (helicopter companies, boats, tourist cottages, tourist stations) and by private citizens and authorities. The organizations taking part in running an economic association for the test bed (the Producers) will in some cases be involved in running the DTN based internet access also. The test bed will reply on infrastructure in form of N4C DTN based Internet access. The Producers competencies (citizens in the area but also researchers and similar staff at the LTU, including the two municipal development agencies in Gällivare and Jokkmokk). Providers and Producers will set up and administrate the Economic Association for the test bed in Swedish Lapland. Monetary consequences of the test bed will be limited to staff for marketing of the test bed. Other costs will be carried by the tester. A business model called service support seller (Hecker´s) will be tested. Revenue anticipated from test service but also complementary service such as transport, hotels, restaurants etc. Full cost principle. Sales of complementary services will be factured by the seller.

8/12

Test of business and governing model for Test Bed Slovenia Based on: Osterwalder, A, Pigneur, Y, Tucci, C, (2005) and Ostrom, (1990)

Pillar

Product Customer Interface

Business Model /Building Block Value Proposition

Description

Target Customer /User Distribution Channel

Universities, SMEs, and ICT industry. MEIS present clients from nuclear industry and clients from new business ideas that will be developed.

Partner Network

MEIS and other Producers have defined roles in the partnership fo r the

Cost Structure Revenue Model

Monetary consequences of the test bed will be limited.

Test Bed services with DTN based internet access as infrastructure.

Test Bed services will be distributed by MEIS and other SMEs and by authorities in the area. Relationship MEIS will build relationship will industry and SMEs interested to take part running the test bed services. Infrastructure Value The test bed will reply on infrastructure in form of DTN based internet Management Configuration access. MEIS and other SMEs who are willing to run the infrastructure or to give service to the testers will form an association similar to the one in Swedish Lapland. Core MEIS will need other SMEs competencies (i.e. companies offering Competency complementary services in form of transport, hotels etc. in the area but also technical and similar staff) to run the test bed.

Financial Aspects

test bed in Slovenia. A business model called service support seller (Hecker´s) will be tested. Revenues are anticipated from test service but also from complementary service such as transport, hotels, etc. Full cost principle. Sales of complementary services will be factured by the seller.

9/12

Applications by MEIS

Type

Meteorological station - Meteorological station Symbinode gamma AMS-DTN with extended autodose rate measuring nomy for critical device - u-garamo situations measurements AMScs

Offering

Stand alone meteorological station with DTN connectivity for estimation of wind power capacity.

Estimation of the possible air pollution dispersion in the cases of accidents at nuclear power stations with releases into the atmosphere.

A simple to use gamma dose rate measuring device for personal use.

Relationship

MEIS will build relationships with consultant companies in wind power development industry.

MEIS business partner is the Nuclear Power Plant in Slovenia. New relationships will be developed.

This service can be sold on line when a USB (or similar device) is developed.

Revenue model

This application will generate revenues to update and maintain the DTN based internet access in Slovenia and to the private company MEIS.

This development is already out for sale by MEIS. The application generates revenues to update and maintain the DTN based internet access in Slovenia and to MEIS.

This application will generate revenues to update and maintain the DTN based internet access in Slovenia and to MEIS.

N4C / MEIS´ Results: •Three devices with services that have been tested with good forecasts. One service is being rolled-out! •The Meteorological station - AMS-DTN contributes to reductions in CO2. •The Meteorological station with extended autonomy for critical situations measurements AMScs contributes to better surveillance of nuclear power plants.

10/12

Apps for Hikers PDA Offering Apps for Hikers use a PDA and is developed for usage in a CCR. The use

Norut´s Apps for Hiker´s PDA for Communication Challenged Regions (CCR)

cases that have been developed and tested are: · DTN based E-mail and Web caching · Auto Discovery: · Geoblog · Send Message with own location · Maps for own location: Downloading maps of the area around one‟s own location. · Point Of Interest (POI) suitable for CCRs: Location of nearest medical services (medical, physical or heart starter) Location of hostels Location of food stores and petrol stations Location of best site for cloudberries, fishing sites etc. Where to get a wash / shower / bath Other issues of location privacy [EFF], security, battery life time, wireless range and ad hoc connections, all very important for the functionality and usefulness of the applications. Relation- The Apps for Hiker's PDA are designed for people working / staying in ship Scandinaivan Artic area. Users are: 1. Professional users (Business): Reindeer herders, tourist guides and forest workers 2. Professional users (Government): Nature Park Rangers and the police 3. Private users (Private citizens): Tourists who are hiking in the area Revenue Two alternative is being considered for the Apps for CCR model 1. To sell it 2. To give it away Sales of the · · · · ·

Apps for Hikers PDA to can go via Companies for outdoor equipment Sami organizations Mountain Cabin and/or Mountain Station Personnel Forest industry Nature park rangers and police force

11/12

Application Herdeview by Tannak Herdview by Tannak

N4C / Tannak´s Results •Trials on using the DTN by a small SME with little ICT experience •Development of models for collaboration with future users

Offering

Products, systems and consulting services for identification and tracking of free-ranging livestock and wild animals, objects, people. The aim is to substantially improve the efficiency and profitability of the livestock owner, and the business value. The USP with Herdview is to radically reduce the use of snowmobiles, motorcycles and helicopters. The technology does not only reduce energy use by motor vehicles and reduced damage to the environment but also contribute to reduction of workplace injuries. This is important because herders are the most injured among workers in Sweden today. The technology can also be used to yield better meat and increase the profitability of reindeer herding.

Relationship

Both Ms. Susan Spik and Ms. Karin Kuoljok are Sami women who have grown up in traditional reindeer-herding families and themselves worked as a reindeer herder in more than 20 years in Sirges village. Susan Spik and Karin Kuoljok have a strong interest in the Sami culture and in the development of a special part of their heritage - the reindeer. Through their N4C work, they have developed their product 'Herdview by Tannak „.Herdview is done in close collaboration with future users.

Revenue model

Traditional pricing /revenue model for Tannak. When the application is used on the DTN based internet access it will generate revenues to update and maintain the DTN based internet access in Swedish Lapland.

12/12

PUBLIC DISSEMINATION

N4C round the world

1/4

N4C DISSEMINATION Outcome in short Scientific Publications

(per 25/03/2011) 13 articles for peer-reviewed scientific publications 5 papers for non peer reviewed scientific publication presented in scientific events such as workshops, conferences, symposia, etc.

Internet draft Probabilistic Routing Protocol for Intermittently Connected Networks draft-irtf-dtnrg-prophet-09 Internet Research Task Force (IRTF) Internet-Draft International work shops in Luleå, Coimbra, Tromsø

Participation in meetings •

• Arctic Council Swedish Governmental Agency for Innovation Systems (Vinnova) • Broadband in Gällivare • The Equal Opportunities Delegation of Norrbotten County - at LTU

FIRE events and work

in Brussels, Stockholm, Gent etc.

Local Work shops

In Slovenia and Sweden, Jokkmokk

N4C Web site (www.n4c.eu)

with 350-400 visitors per month from more than 35 countries (EU, Brazil, USA, Iran, China, India etc.)

Produced online material

Newsletters 5 English issues and 3 Portuguese issues

2/4

N4C PARTNERS EXPLOITATION PLANS When you took part in N4C work was your intention then (in 2008) to

N4C was not started with intention to develop new companies / services that would generate profit. Yet, TCD have developed a company: Tolerant Networks Ltd..

develop a computational tool or a service or other type of innovation that you can? (Check as many as apply!)

7

5

3 2

Seven partners (SMEs and research institutes) had the intention to use what was developed in the business they already have. The “give away” strand as an “OSS” or just as a “give away to other communities” is very strong among all partners (universities, research institutes, industry and SMEs).

2 1

1

0 Start a new Exploit in the Make Give away to Use in your Use as a My intention Other state business and business you available for other academic showcase in was not to what exploit (make already have other as communities career and your exploit profit on) open source get academic consultant anything software credit for service

These results coincides with the large interest N4C has generated from different countries Malaysia, Brazil, Mongolia, Alaska to mention some. (See international interest in N4C web site.) The aim to use results in “academic career and get academic credit for” was very strong for four partners from academia. For more details: see Annex Exploitation Plan

3/4

N4C EXPLOITATION AND DISSEMINATION

Partners have taken part in developing new projects -

A consortium has been set up by IPN to make a proposal to FP7 for a project in Amazonas, Brazil. Participating were IPN, LTU, PLAB and TCD.

In December 2010 a new project, eLearningDTN, started at LTU. In the on-line course, How to build and implement DTN based internet access with gender and cultural sensitive approach the efforts from N4C project will be transferred to new audiences. Groups that will be targeted are SMEs, NGOs, schools etc. , and other who need communication and are interested in using DTN. Partner in the project are LTU, Sweden, Tolerant Networks Ltd., Ireland, MEIS, Slovenia, IPN, Portugal, and PLAB Sweden.

4/4

PRoPHET DTN Routing PRoPHET => Probabilistic Routing Protocol using History of Encounters and Transitivity

Introduction and a Little History! Concept invented by Avri Doria and Anders Lindgren for the SNC (Sámi Network Connectivity) project in 2002. Designed for DTNs where there is no fixed topology or schedule. All data forwarding happens at opportunistic encounters between nodes. Patterns in the mobility are used to improve the use of resources as compared with Epidemic Routing to which it is related. Developed and implemented during SNC and SNC+1, assisted by Samo Grašič who built the Prophet DTN infrastructure software used in N4C. Looking to have the latest version of the specification published as an Experimental RFC under the auspices of the IRTF DTN research group.

13 April 2011

1/9

Folly Consulting

Epidemic Routing Simple but resource expensive scheme for getting DTN bundles to every encountered node – and hence necessarily to the intended destination – eventually, provided that bundle lifetime is adequate Algorithm: Whenever an opportunistic encounter occurs exchange any bundles with the encountered node that the node doesn’t already possess – Result: If all goes well both nodes have the union of the sets of bundles they both had before the encounter

Advantage: Guaranteed to find the optimum path… .. because it tries every path Disadvantages: – This is all very well if nodes have infinite storage capacity – Practical encounters may not last long enough to complete the exchange .. And it would be good to prioritize transfers that expedite delivery

13 April 2011

2/9

Folly Consulting

PRoPHET improves Epidemic In a real network involving human activity, where mobile nodes are controlled or carried by humans: – Mobility is not truly random as in the sense of ‘white noise’ – Human activity imposes patterns on the mobility – The mobility has a characteristic time interval that depends on the sort of activity we are talking about • Typically this is much longer than the sort of delays that are normally found in the connected Internet • On the order of hours or days in many circumstances such as N4C’s test beds

PRoPHET extracts the essence of the mobility pattern of nodes – abstracts ‘History of Encounters’ into delivery predictability parameter – parameter is used to prune the Epidemic routing paths – aim to avoid sending bundles only on paths that have a low probability of reaching the bundle’s destination

13 April 2011

3/9

Folly Consulting

Delivering a Bundle This sequence of diagrams shows how a set of Wi-Fi equipped nodes moving in an area that is much bigger than the range of Wi-Fi might deliver a bundle Time t: A sends message for D

Time t+dt: A meets B – good path

D

D

Time t+2dt: B meets E – bad path C

D

B B

C A

E

Time t+5dt: C meets D – delivery!

C A

E

A

E

Message passed from A to B

Message not exchanged

Time t+4dt: B meets A – good path

Time t+3dt: B meets C – good path

D

E

B

D

C

D

C B

C

A

B

A

B

Message passed from C to D

No exchange – both have message Node with Wi-Fi range

13 April 2011

A

E

A

Message from A to D 4/9

E

Message passed from B to C Copy of message

Folly Consulting

Modelling the Mobility Pattern Every node maintains a set of Delivery Predictabilities (DPs) for nodes it has encountered (reasonably) recently –

DP for node B stored in node A: PA(B)

–

DP before/after encounter: PA(B)old / PA(B)new

The DPs evolve over time as nodes encounter each other, thus… Basic DP Evolution Equations in node A when it meets node B 1. Direct encounter: DP for encountered node increases at each encounter •

PA(B)new = 0.5 [on first encounter – when PA(B)old is 0] •

•

don’t know whether there will be more meetings so ‘hedge our bets’..

PA(B)new = PA(B)old + (1 - PA(B)old) * Pencounter [subsequent encounters]

1. Decay over time: All DPs are decreased if the node hasn’t been encountered •

PA(B)new = PA(B)old * γK [K is number of time units since last decay] •

If PA(B) gets very small, set it to 0 and treat next encounter (if any) as first

1. Transitive rule: If B is a good path to C and A meets B frequently then nodes that meet A might want to give messages for C to node A •

PA(C)new = max(PA(C)old, PB(C), * PA(B)new * β) [β is a constant] •

Encountered node (B) sends its set of DPs to A for use in the Transitive Rule

Folly Consulting

The real world is a bit more complicated (Pencounter is not a constant) 13 April 2011

5/9

PRoPHET Protocol Manages encounters between pairs of nodes. – Effectively a point-to-point protocol

Stage 1: detect a new neighbor – Mechanism depends on link layer and separate discovery protocol – For example, DTN2 has IP and Bluetooth based mechanisms

Stage 2: execute PRoPHET Hello protocol to confirm PRoPHET support – Agree first/second roles between connected pair; exchange identities

Stage 3: information exchange – details on next slide – On completion both nodes have sent and received some bundles according to relative values of updated DPs in the two nodes

Stage 4: extra bundle exchange – If new bundles arrive due to local applications or new encounters determine if they should be sent to connected node(s)

Stage 5: periodically repeat complete information exchange Neighbor disappears: break off connection – may happen at any stage

13 April 2011

6/9

Folly Consulting

PRoPHET Information Exchange PRoPHET Information exchange stage: – Both nodes apply decay equation (2) to own DP sets – Nodes send these sets of DPs to the encountered neighbor – Nodes apply direct encounter and transitive equations combining its own DPs and the set of DPs received from the encountered neighbor – In each node, compare updated local DPs with DPs received from encountered neighbor: • According to policy and current DPs determine if any bundles held by this node should be offered to the encountered node

– Each node sends offers to other node • Node receiving offers decides which bundles to accept according to local policy e.g., will the bundle fit into available storage? is it too old? • Having decided which bundles to accept it sends acceptance responses to offering node posibly modifying the order to suit its own policy

– Offering node then sends accepted bundles to the other node in the order requested by the accepting node

13 April 2011

7/9

Folly Consulting

PRoPHET in N4C The Prophet DTN infrastructure code as developed by Samo Grašič has been extensively used for – The LTU web/email/nsim/podcast experiments in Swedish Lapland. – The MEIS meterological and radiological monitoring experiments in Kočevje, Slovenia.

The PRoPHET routing protocol code in the DTN2 reference implementation has been tested and had limited experimentation in conjunction with the PyMail DTN nomadic email system developed by Folly Consulting and deployed to a limited extent in Swedish Lapland. A new version of PRoPHET has been researched, designed and specified by Elwyn Davies of Folly Consulting. Extensive simulation of the new and older versions of PRoPHET have been carried out by Samo Grašič at LTU to demonstrate some issues with the older version and compare it with the new version. – An academic paper is in preparation

13 April 2011

8/9

Folly Consulting

PRoPHET Resources Current Internet Draft specifying Version 2 of the PRoPHET protocol: http://tools.ietf.org/html/draft-irtf-dtnrg-prophet-09 Source code of Prophet DTN Infrastructure software: http://code.n4c.eu/code/PRoPHET/ Source code and documentation of the PyMail nomadic email system http://code.n4c.eu/code/PyMail/

13 April 2011

9/9

Folly Consulting

DTN Fundamentals DTN => Delay- and Disruption-Tolerant Networks/Networking Why do we need DTN? ‘Conventional’ Internet assumes • short end-to-end delays • fairly stable connections

Protocols (especially TCP) and many applications don’t work well on networks that don’t match these requirements

Often in Communication Challenged Regions (CCRs) neither is satisfied Some DTN Use Cases for ‘stressed’ networking environments

Space: The Interplanetary Internet

N4C CCR Scenario

Terrestrial: Opportunistic connections

Terrestrial: Roughly Scheduled connections

Areas with no wired and minimal wireless comms. Data transfer during encounters according to statistical pattern

Areas with no wired and minimal wireless comms. Data on USB stick, transport by bus

Examples… Comms from Earth – Space Probes Very long delays, unidirectional Scheduled comms opportunities

13 April 2011

1/7

Folly Consulting

Store, Carry, and Forward Scenario

Data Delivery Paradigm

Routing Mechanism

Conventional Internet (IP packet forwarding)

Independently handled packets with immediate forward or discard

Precalculated Routes using topology knowledge (via OSPF, BGP etc)

Email (RFC 5322)

Unitary Transfer of Overlay of configured links complete emails with Store between (fixed) mail and Forward at each servers plus routing intermediate hop configuration

DTN Unitary Transfer of (using RFC 5050 Bundle ‘Bundles’ with Store, Carry Protocol) and Forward at each intermediate hop node; nodes maybe mobile.. Hence ‘carry’.

13 April 2011

2/7

Depends on Use Case: Static Configuration, Schedule Oracle, Delay Tolerant Link State or (Un)Pruned Epidemic

Folly Consulting

RFC 5050 Bundle Protocol Bundles are complete encapsulated units of data delivered in a DTN, including – – – –

The message data, either encrypted or not, according to requirements Addressing and routing control information Descriptive metadata, with extensible scheme in case extra is needed in future Optionally, security control information for encryption and/or integrity protection

Multiple instances of same bundle might exist concurrently in different parts of a DTN Key features of a DTN using RFC 5050 –

Custody-based retransmission for bundles •

– –

Ability to cope with intermittent connectivity Ability to take advantage of • • • •

–

responsibility for delivery of a bundle can be transferred from sender to other willing nodes

Scheduled connectivity Predicted connectivity Opportunistic connectivity (And also) Continuous connectivity

Late binding of endpoint identifiers (EIDs or DTN addresses) to constituent (internet) addresses •

The path a bundle takes through the network is not determined at the point it leaves the sender (unlike Internet)

13 April 2011

3/7

Folly Consulting

Routing for DTNs Mechanism

Description

Static Routing

Preconfigured, table based – specify either next hop or intermediate waypoint

Delay Tolerant Link Route calculation based on dynamically discovered State Routing (DTLSR) topology. Unlike Internet LSR links can be potential rather than actual provided the overall topology is known in advance. Scheduled (aka Oracle- A central Oracle provides all knowledge of what based) Routing links will be available at what times allowing forwarding hops to be calculated and scheduled. Epidemic Routing

Every bundle carried in a node is passed to every node that links to the carrier. Highly resource intensive!

PRoPHET Routing

A form of pruned epidemic routing. Statistical patterns in the mobility of nodes are used to reduce the demands on resoutces Folly Consulting 4/7

13 April 2011

DTN Addressing Addresses for a DTN use a Uniform Resource Identifer (URI) scheme Basic scheme is ‘dtn:’ – Combines a node ‘identifier’ (similar to a DNS name), and – Path information used to route bundles to applications. – May also contain more complex information to support content-based or service based routing/delivery.

The dtn: scheme is still under development There is also an ‘ion:’ scheme suitable for space based systems and scheduled routing

13 April 2011

5/7

Folly Consulting

RFC 5050 DTN Implementations N4C deployments are based on either: – IRTF DTN Research Group coordinated DTN2 reference implementation, or • Information from http://www.dtnrg.org • Code available from http://code.n4c.eu or http://dtn.sourceforge.net

– Prophet implementation derived from previous SNC project • This implementation is based on an earlier version of the bundle protocol and is not compatible with DTN2. • Code available from http://code.n4c.eu

Other implementations exist. More information can be obtained from the DTN Research Group web site http://www.dtnrg.org

13 April 2011

6/7

Folly Consulting

More Information on DTN ‘DTN - The State of the Art’: N4C Deliverable D2.1 contains information about the history development and current experimentation with DTN. Downloadable from http://www.n4c.eu/Download/n4c-wp2-012-state-of-the-art-101.pdf ‘Delay-Tolerant Network Architecture’: RFC 4838 describes the overall architecture of DTN. Downloadable from http://tools.ietf.org/html/rfc4838 ‘The Bundle Protocol’: RFC 5050 describes the basic bundle protocol used in N4C and elsewhere. Downloadable from http://tools.ietf.org/html/rfc5050 ‘Functional Specification for DTN Infrastructure Software’: N4C Deliverable D2.2 describes the DTN2 reference implementation. Downloadable from n4c-wp2-023-dtn-infrastructure-fs-12.pdf

13 April 2011

7/7

Folly Consulting