IXSI - Interface for X-Sharing Information

Wolfgang Kluth, Markus C. Beutel, Sevket G¨okay,

Karl-Heinz Krempels, Christian Samsel and Christoph Terwelp

Information Systems, RWTH Aachen University, Aachen, Germany

Keywords:

Travel Information, Intermodal Transportation, System Integration, Carsharing.

Abstract:

The increasing demand for mobility, especially for individual transport, leads to more pollution, congested

cities and shortage of parking. New ways of mobility could mitigate these issues. Unfortunately, such forms

of mobility, i.e. carsharing, are usually isolated services, because of the missing integration with other mobility

modes. Our aim is to offer a combination of heterogeneous services on a single platform. We argue that such

an integration allows optimal offers and higher usability and therefore results with a higher acceptance among

travelers. Embedding a vehicle rental system into a travel information system information-wise is a step

forward. For this purpose, we developed an interface for x-sharing information, short IXSI, specialized in

connecting vehicle rental systems with a travel information system. IXSI is an XML-based, B2B interface

with functions for, e.g., exchanging basic vehicle data and price information. The interface enables travel

information systems to perform bookings of carsharing and bikesharing vehicles and therefore allows the

customer to use traditional public transport services as well as rental services seamlessly. Furthermore, we

brieﬂy present our IXSI implementations on travel information and vehicle rental system side.

1 INTRODUCTION

Carsharing is a growing market worldwide (Hayashi

et al., 2014). In the past decade, the number of shared

vehicles, as well as the number of members using

carsharing have increased rapidly (Shaheen, 2012).

Forecasts predict an extensive growth of carsharing

users from 700,000 in 2011 up to 15 million in 2020

in Europe (Statista GmbH, 2015). Nevertheless, car-

sharing is usually seen as a supplement to traditional

public transportation services like trains or under-

ground railways in order to improve service coverage

(Huwer, 2004). That is to say, a customer could travel

to a different city by train and switch to a carshar-

ing vehicle to reach his or her ﬁnal destination which

might not be reachable via public transport.

Although such a journey is already possible today

by manually booking a train ticket as well as a car-

sharing service, a single point of contact for planning,

booking and assisting a trip would greatly simplify

the journey. To render such a single point of con-

tact possible, system integration is required. Unfortu-

nately, due to the heterogeneity of different mobility

modes, e.g., train schedule vs. individual transport,

system integration efforts are demanding.

For the following remarks, we deﬁne a travel in-

formation system (TIS) as a system covering (travel)

data integration, routing, price calculation, booking,

and the presentation of information through a user

interface. This system is supposed to be the SPOC

for all travel related matters. Assisting the customer

while traveling is also conceivable. Moreover, we de-

ﬁne a vehicle rental system (VRS) as an information

system responsible for administration and accounting

of rental vehicles and customers. Rental vehicles in

this context can be cars (with conventional engine or

electric) or bikes (with or without electric drive).

Hence, we contributean approach, interfacing TIS

and VRS under consideration of different procedures,

functionalities and requirements.

This work is structured as follows. First, Sec-

tion 2 summarizes related work in the ﬁeld of interest.

As a focal point, Section 3 describes the contributed

approach and illustrates functionalities by means of

relevant use cases. Afterwards, Section 4 covers the

actual information exchange in detail and Section 5

presents our implementation efforts. Finally, Sec-

tion 6 reﬂects the approach and describes future work.

2 RELATED WORK

In the particular area of travel information systems,

293

Kluth W., C. Beutel M., Gökay S., Krempels K., Samsel C. and Terwelp C..

IXSI - Interface for X-Sharing Information.

DOI: 10.5220/0005506102930298

In Proceedings of the 11th International Conference on Web Information Systems and Technologies (WEBIST-2015), pages 293-298

ISBN: 978-989-758-106-9

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

system and data integration is a prominent research

ﬁeld (Figueiredo et al., 2001).

(Fluhr and Lutz, 2011) deﬁne special use case

types for electric vehicles and underline their impor-

tance for the development of comprehensiveinforma-

tion system architectures. Similarly, our work starts

out from outlining the use cases.

(Barth et al., 2002) focus on wireless communi-

cation of shared vehicle systems and describe a data

communication architecture for sharing systems with

multiple stations. (Beutel et al., 2014) propose an ar-

chitecture for comprehensive travel information sys-

tems and touch on the issue of interfacing heteroge-

neous mobility modes. Thereby, methods of data in-

tegration are addressed.

Furthermore, (Doshi, S., 2010) focuses on the de-

velopment of a multimodal traveler information sys-

tem and describes an interface with a limited amount

of functionalities.

3 USE CASES AND APPROACH

Integration of different mobility providersinto one in-

formation system which provides a central interface

to the user is challenging. Current methods which in-

volve redirecting users from the TIS to the VRS for

further information and booking services, lowers the

usability and quality of the process.

Therefore, a protocol is needed which allows a

seamless integration of one or more VRS into a sin-

gle TIS which handles vehicle information, bookings,

and price information.

3.1 Use Case Evolution

To investigate possible effects, two use-cases are pre-

sented, with and without VRS integration, taking into

consideration the roles represented in Figure 1.

TIS Operator

Travel

Information

System

VRS Operator

Vehicle Sharing

System

User TIS

IXSI

Figure 1: Overview Roles.

Without Integration

Today, Jack has a business appointment with a new

customer to discuss a project in Cologne later that

day. At the moment, he is sitting in his ofﬁce in

Aachen, so he needs to plan a trip from Aachen to

Cologne. For this purpose, he uses a TIS to request

several travel chains. The results represent solutions,

using public transport, but no VRS options. To get

further options, e.g. car sharing, he has to explicitly

query a VRS to get information about the availability

of nearby stations and cars. Jack has decided to take

the car sharing option, despite the fact that he has to

walk one kilometer to the next car station. He stops

further efforts to ﬁnd a better solution, because com-

bining public transport and car sharing is too complex

and time-consuming.

With Integration

Jack’s colleague, Linda is in the same situation, but

her appointment is one day later. She has access to a

novel TIS with a full integration of the VRS. As input

for her request she is using her ofﬁce address as the

starting point, the customer’s address as the destina-

tion, the arrival, and return time. As result she gets

a list of trips with public transportation and car shar-

ing, separated and combined. She decides to take the

route which involves traveling by bus to the carshar-

ing station for the ﬁrst kilometer. From there she is

going to drive the carsharing vehicle the rest of the

way. She is not worried about her plans, because the

TIS checked implicitly the availability of the car and

in the next step she can book it directly without in-

volving the VRS.

Interlinking Between TIS & VRS

Considering both use cases, the integration of the

VRS into the TIS shows signiﬁcant improvements

of usability. One fundamental piece is the integra-

tion of the VRS into the TIS, which covers all neces-

sary functionalities (e.g., availability query and vehi-

cle booking). For this purpose, IXSI (Interface for X-

Sharing Information), which connects the VRS with

the TIS, was developed, In the following section the

functionality and general structure of IXSI are pre-

sented.

3.2 Hierarchy Model

The hierarchy model of IXSI describes different qual-

ities of information coupling based on service groups.

Furthermore, it contains recommendation for various

levels of implementation.

The interlinking between TIS and VRS needs at

least the implementation of Service 1 (static data) by

both systems. Figure 2 shows the dependencies be-

tween services.

Service 1 (static data) exchanges information about

providers and static data of booking targets (see

Subsection 4.1). For example, sharing stations

(VRS) can be presented in a TIS.

WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies

294

Service 1

Static Data

Service 3

Availability

Subscription

Service 2

Availability

Query

Service 4

Booking

Service 6

Price

Information

Service 5

Booking

Subscription

Figure 2: IXSI Service Groups dependencies.

Service 2 (availability query) provides queries for

availability information. When a customer is in-

terested in an itinerary containing a speciﬁc vehi-

cle, the actual availability times of the respective

booking target are requested to ensure the vehicle

is available.

Service 3 (availability subscription) depends on

Service 1 and is responsible for an asynchronous

exchange of availability information. The TIS

can subscribe to booking targets in order to avoid

multiple requests. After subscribing to a set of

booking targets, the VRS continuously informs

the TIS about the availability status. This service

is an addition to Service 2 which allows the

information exchange to scale up by preventing

necessary messages and also potentially increases

the interactivity of the service by preventing

delays.

Service 4 (booking) depends on Service 1 and al-

lows vehicle booking, and changing or cancella-

tion of those by the TIS on behalf of the (VRS)

customer.

Service 5 (booking subscription) depends on Ser-

vice 4 and provides a subscription to booking up-

dates. The TIS can subscribe to existing bookings

in order to receive updates from the VRS if the

booking has changed (i.e., vehicle is not available

anymore).

Service 6 (price information) depends on Service 1

and provides price information of rental services.

Based on the starting point, destination, start time,

and end time, the TIS can request price informa-

tion (individualized for a customer or not) from

the VRS, which responds with a tentative price

and if necessary with individual items.

4 INFORMATION EXCHANGE

This section gives a brief overview about how the in-

formation exchange between travel information sys-

tem and vehicle sharing system is actually conducted.

4.1 Data Model

To understand the information ﬂow, understanding of

how real world properties are modeled is required.

This work does not contain a complete list of all types

and attributes, but instead outlines the relation be-

tween the exchanged objects among themselves and

the real world. Please refer to the speciﬁcation

for

more details.

Place describes a location where rental vehicles are

available for rent or to return. Typical examples

are bike sharing or car sharing stations.

Booking Target is an abstract representation of one

or more vehicles which can be rented. It usually

denotes a group of similar characteristics. The

reason for the abstraction is that the provider can

replace a vehicle with another at any time.

Availability is a time span during which a booking

target is free for booking.

Booking is a binding reservation of a booking target

for a speciﬁc time span.

4.2 Information Flow

The following UML-like sequence diagrams give an

overview of a possible information exchange using

IXSI. In the depicted information exchange, the user

inquires the TIS for a travel information, books a

speciﬁc itinerary and subscribes to notiﬁcations. As

shown in the different diagrams, static information

is exchanged as initialization (i.e., places), asyn-

chronously to inform about updates (i.e., booking tar-

gets), and synchronously (i.e., bookings). Although

complex, these different ways of exchanging data, de-

pending on the purpose, allow efﬁcient as well as low-

delay communication.

As initialization the VRS conveys the booking targets

to the TIS and TIS subscribes to changes of those (see

Figure 3). This happens proactively without any user

involvement. As soon as a booking target changes,

the VRS informs the TIS asynchronously.

For the actual travel information (Figure 4), the

user makes a door to door request, using, e.g., a

smartphone. For the respective itineraries, multiple

https://github.com/RWTH-i5-IDSG/ixsi

IXSI-InterfaceforX-SharingInformation

295

:User

:TIS :VRS

Request

List of Booking Targets

Subscription Booking Targets

Updates of Booking Targets

Booking Targets ExchangeBooking Targets Exchange

Service 1

Figure 3: Interaction Sequence: Booking Target Exchange.

:User

:TIS :VRS

Travel Information Inquiry

Availability Request: Booking Targets, Time Period

Contemplable Booking Targets

Availability Request

Service 2

Price Information: Booking Targets, Time Period

Respective Prices

Price Information

Service 6

Session (anonymous)Session (anonymous)

Base Service A

Intinaries

Travel Information InquiryTravel Information Inquiry

Figure 4: Interaction Sequence: Travel Information.

rental vehicles are possible, for which the TIS syn-

chronously checks the availability. Furthermore, the

prices are obtained. All in all, the TIS returns the pos-

sible itineraries and overall prices to the user device.

In the next step (Figure 5), the user selects a suitable

itinerary and wants to book it. He logs into the mobile

app using his personal credentials. These credentials

are passed by the TIS to the VRS to create a session

(usage of an access token is omitted for simplicity).

During the session, the actual booking for the vehicle

is conducted. To book a vehicle, the TIS requests a

booking using a booking target reference and a time

span proposal, which is answered with a booking con-

ﬁrmation and the actual booked time. Finally, the TIS

informs the user about the successful booking.

Optionally, the TIS can subscribe to booking changes

(e.g., cancellations). This is used to inform the user

about such events and to automatically offer alterna-

tives (Figure 6).

:User

:TIS :VRS

Booking Target Reference, Time Period Proposal

Booking Target Reference, Time Period

Vehicle Booking

Service 4

SessionSession

Base Service A

Booking Conﬁrmation

BookingBooking

Figure 5: Interaction Sequence: Booking.

:User

:TIS :VRS

Booking Reference

Booking SubscriptionBooking Subscription

Service 5

Booking Updates

Booking Update

Travel Monitoring

Service 5

Figure 6: Interaction Sequence: Travel Monitoring.

4.3 Message Format, Message Coding

and Transport

Messages between TIS and VRS are transferred as

XML according to a standardized XML Schema.

Messages belong to one of ﬁve archtypes, Re-

quest, Response, SubscriptionRequest, Subscription-

Response and SubscriptionMessage. Request and Re-

sponse form the identically named information ex-

change pattern, whereas the latter three allow a asyn-

chronous data exchange as depicted in, i.e., Figure 3.

As the interface allows asynchronous subscription

besides the usual request/response scheme, usage of

the WebSocket protocol instead of plain HTTP is rec-

ommended. WebSockets allow persistent connections

between both systems and a bidirectional message ex-

change. See (Pimentel and Nickerson, 2012) for a

comparison of HTTP and WebSockets.

4.4 Security and Encryption

IXSI is a B2B interface and therefore does not in-

clude an internal authentication mechanism. Instead

the communication partners are advised to use exist-

ing authentication mechanisms such as SSL certiﬁ-

WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies

296

Mobility Broker

SIRI

others

Train

Bus

Metro

Carsharing

BikeMan

(Bikesharing)

Parking

Carpooling

IXSI

Travel Information System

Vehicle Rental

System

Mobility Services Interfaces

Customer

Station

Figure 7: Overview Mobility Broker Architecture.

cates or a virtual private network (VPN) authen-

tication.

To ensure conﬁdentiality of the transferred data,

encryption is required. Both the SSL/TLS security

layer and VPN are suitable.

5 IMPLEMENTATION

In the previous sections we presented the purpose,

structure and processes of IXSI. In addition to the the-

oretical work, we implemented IXSI for our research

project Mobility Broker (Beutel et al., 2014). Mobil-

ity Broker is an information system which integrates

multiple heterogeneous mobility services including -

but not limited to - buses, bikesharing and carsharing,

into one joint platform. As shown in Figure 7 Mobil-

ity Broker communicates with service providers using

different proprietary and standardized B2B interfaces

(IXSI being one of them) and with customers (respec-

tively their applications) using open APIs for web and

mobile application development.

From a technical standpoint, it is a Java EE 7 ap-

plication consisting of many modules and running on

a WildFly

application server. The IXSI module is a

WebSocket client and leveragesthe Java API for Web-

Socket (JSR 356) implementedby the Undertow

web

server, which is part of WildFly.

As for the VRS, we chose to extend our open

http://wildﬂy.org/

http://undertow.io/

source tool BikeMan

with the IXSI implementation.

BikeMan is a station-ﬂexible e-bike/pedelec sharing

system which enables the management of stations,

pedelecs and customers. Moreover, it allows to mon-

itor pedelec rentals. BikeMan is a Java web appli-

cation based on Spring Boot

and published under

Apache License, Version 2. In the interaction with

Mobility Broker,BikeMan acts as a WebSocket server

and uses Spring’s own WebSocket API which is JSR

356 compatible.

Since WebSocket is a low-level transport protocol,

it does not provide the following messaging conve-

niences out of the box:

1. RPC-style synchronous messaging: WebSocket

is asynchronous and only covers exchanging of

data or text frames, but not communication with

request-response pattern.

2. Message routing: It does not interfere with the in-

terpretation of the content, namely the determina-

tion of the message type and which procedure to

subsequently invoke.

3. Error handling: It does not offer any mechanism

for reliable message deliveryor fault-tolerant con-

nectivity.

Therefore, we had to tackle some challenges on both

sides and build our own abstractions to deal with the

high-level interface that is IXSI.

https://github.com/RWTH-i5-IDSG/BikeMan

http://projects.spring.io/spring-boot/

IXSI-InterfaceforX-SharingInformation

297

6 DISCUSSION AND OUTLOOK

The aim of this paper was to present our approach of

integrating sharing/rental vehicles into an intermodal

journey by connecting the related IT systems. To

do so, an asynchronous, XML-based interface was

created. Our presented implementation shows that

technology-wise an integration is complex, but pos-

sible. Unfortunately, the technical implementation

is only the ﬁrst step of offering an integrated mo-

bility service to citizens. There are still political,

economical and sociological challenges involved and

might even contain show-stoppers, which could po-

tentially render the technical solutions pointless. Fur-

thermore, the interface as well as the implementation

are not in productive use and require more investiga-

tion. Speciﬁcally, performance and stability will be

focal points.

We are currently in the phase of deploying our

prototype to a closed user group to identify further po-

tential problems in technical and other aspects. After

the system operation has been assured, user studies

will commence to quantify the acceptance of a sys-

tem, integrating sharing/rental systems and traditional

public transports.

As a ﬁnal step of development, we plan to stan-

dardize IXSI in close collaboration with a suitable in-

ternational organization.

ACKNOWLEDGEMENTS

IXSI was developed in cooperation by Cantamen

GmbH, HaCon Ing.-Ges. mbH and RWTH Aachen

University for project eConnect Germany. The au-

thors would like to thank Peter von Grumbkow, Dirk

Hillbrecht, Heike Twele and Gerhard Wagner for their

work on IXSI. This work was funded by German

Federal Ministry of Economics and Technology for

projects eConnect Germany (01ME12052) and Mo-

bility Broker (01ME12135A).

REFERENCES

Barth, M., Xue, L., Chen, Y., and Todd, M. (2002). A hy-

brid communication architecture for intelligent shared

vehicle systems. IEEE Intelligent Vehicle Symposium

(Vol. 2), pages 557–563.

Beutel, M. C., Goekay, S., Kluth, W., Krempels, K.-H., Ter-

welp, C., and Samsel, C. (2014). Product Oriented In-

tegration of Heterogeneous Mobility Services. IEEE

17th International Conference on Intelligent Trans-

portation Systems (ITSC), pages 1529 – 1534.

Doshi, S. (2010). Designing a Multimodal Traveler Infor-

mation Platform for Urban Transportation.

Figueiredo, L., Jesus, I., Tenreiro Machado, J., Rui Ferreira,

J., and Martins de Carvalho, J. (2001). Towards the

Development of Intelligent Transportation Systems.

IEEE International Conference on Intelligent Trans-

portation Systems (ITSC).

Fluhr, J. and Lutz, T. (2011). Use Case Types for Communi-

cation with and for Electric Vehicles (EV). 2011 17th

International Conference on Concurrent Enterprising

(ICE), pages 1–6.

Hayashi, T., Hidaka, K., and Teshima, S. (2014). Car-

sharing and IT Enabled Services. 2014 Annual SRII

Global Conference, (Fig 1):274–280.

Huwer, U. (2004). Public transport and csar-sharingbeneﬁts

and effects of combined services. Transport Policy

Vol. 11, pages 77–87.

Pimentel, V. and Nickerson, B. (2012). Communicating

and displaying real-time data with websocket. Inter-

net Computing, IEEE, 16(4):45–53.

Shaheen, S. (2012). Understanding from Shared-Use Mo-

bility Research.

Statista GmbH (2015). Anzahl der Carsharing-Nutzer in

Europa in den Jahren 2011 und 2020 (in Millionen).

WEBIST2015-11thInternationalConferenceonWebInformationSystemsandTechnologies

298