IEEE Double Column

Service Integration Platform based on TINA 3-tier Model and Interfaces

Junichi Toyouchi*, Motohisa Funabashi*, and Linda Strick**

 * Systems Development Laboratory, Hitachi, Ltd., 1099 Ohzenji, Asao, Kawasaki 215-0013, Japan, {funa, toyouchi}@sdl.hitachi.co.jp ** GMD FOKUS Berlin, Kaiserin-Augusta-Allee 31, D-10589 Berlin, Germany, [email protected]

Abstract natural language translation. In February 1999, Presi- dent’s Information Technology Advisory Committee Electronic commerce system will be combined and (PITAC) issued a report titled “Information Technology collaborated over the network into single services just Research: Investing in Our Future”[2] in the US. The like idea of global computing within very short time. In report mainly discusses in three areas, namely software, future, the services will be self-organized for pursuing scalable information infrastructure, and high end com- human needs in communities and cultures. According to puting. For the software, importance to work on soft- this bold vision, Japanese and European parties are de- ware engineering, HCI, and information management is veloping service integration platform collaboratively stressed. For the scalable information infrastructure, based on autonomous decentralized service system R&D for bandwidth, modeling of dynamic behavior of (ADSS) architecture influenced by TINA business model network, and middleware for scalability and survivabili- and functionalities. ty are requested. For the high-end computing, innova- tive architecture, software for high-end computing, and 1. Introduction peta-flops machines are main focuses for R&D. Expectations for the information and tele-communication revolution are pervasive. In business domain, the It is true that we need much more bandwidth of com- Internet and other telecommunication services are antic- munication in future, but so far not so much has been ipated to drive the market to open for creating tremen- discussed on the middleware for information sharing dous business opportunities as well as to permit man- over the network even though it plays a crucial role for agers to make real-time decisions for extending their realizing the expected revolution. Present middleware business chances. In social domain, the revolution is ex- technologies for information sharing stay at very primi- pected to reduce environmental loads and consumption tive capability such as email, Web, FTP, etc. These tech- of natural resources. It is also expected to leverage mu- nologies actually played one of the important roles tual understanding among individuals over the world, opening the revolution, and still keep the usefulness and forming communities, and fostering knowledge and cul- importance. However, presently we need much more tures. sophisticated information sharing capability for execut- ing complex business processes and social interactions According to these expectations, many countries are over the network and these technologies have not been discussing policies to steer the development of IT and well elaborated. social systems for the coming age. For example, in March 2000, Ministry of Posts and Telecommunica- Projects for global computing such as Ninf [3] are tions, Japan has issued a report named “Vision of Infor- presently running which try to collaboratively use mation and Telecommunication in 21st Century” -IT servers distributed over the Internet. The main goal of Japan for All-” [1]. In terms of technological develop- these projects is to extract high performance computing ment, the report points out necessity of developing di- power from servers distributed over the world. To some versified access networks, peta-bps class backbone net- extent, the idea of this collaborative use of distributed works, smart clients and environment for communica- servers would provide technological hints for information and broadcast integration, as well as technologies tion sharing needs, but it is not enough. Formatted for freeing access barriers such as voice recognition and As exceptional efforts for sophisticated information Second wave: Mega-media, which is initiated by sharing, TINA [4], ADSS (Autonomous Decentralized 6Mbps of one directional communication, permits me- Service Systems)[5] that is influenced by TINA, and E- dia bazaar (market) where more than 10 thousands soft- speak [6] can be mentioned. TINA and ADSS are partial ware and contents venders are expected to participate in ancestors of the present research that will be shown in Japan. later section in this paper. E-speak [7] appeared recently stemming from idea of global computing. It focuses Third wave: Mega-contents, which are initiated by mainly on just collaboration mechanism and business 10Mbps of bi-directional communication, bring explo- interactions among the participants are treated as appli- sive big content flows among SOHO, communities, and cation matter in contrast to TINA and ADSS. home.

The objective of this paper is two fold. One is to In contrast to the three business waves, Prof. Masuda present our idea for sophisticated information sharing views the revolution in a different way [8]. He claims and service integration needs in which chaining of non- that ECs will include Electronic Commerce as well as deterministic service provisioning plays an essential Electronic Communities and Cultures. Each EC is char- role, and the other is to propose international joint ef- acterized as the follows; forts assuring inter-operability among related functions. The composition of this paper is as the follows; in Sec- Electronic Commerce: pursuing economic rational- tion 2, future visions and expectations for the coming ity by effective services and giving economic founda- information society are reviewed for defining the re- tion for developing and maintaining society. quirements for the information sharing systems. In Sec- tion 3, our concept, architecture, and essential parts of Electronic Communities: providing legitimate pub- sophisticated information sharing and service integra- lic services, including such as care management infor- tion is presented, and in Section 4 an international col- mation systems, government service systems, and learn- laboration developing the middleware with a prototype ing systems. In Japan, system solutions for these areas system are presented. will produce $890B and 4.8 million employments in 2010. 2. Scope Electronic Cultures: supporting intellectual and 2.1. Emerging three business waves knowledge based activities for cultural legitimacy such as digital libraries. In Japan, transactions for contents In every country, the network revolution is expected will amount to $150B in 2010. to bring at first B2B and B2C business. In Japan, it is predicted that market size of B2B and B2C will grow to $647B and $30B respectively in 2003. For B2C, contri- 2.2. Typical applications bution of mobile Internet systems will be prominent, In order to clarify the requirements for sophisticated permitting access to the Internet including payment information sharing and service integration, let us ex- through the mobile phone. For example, present number amine typical future applications in both business and of subscribers of i-mode by NTT DoCoMo has exceed- social domains. ed 5 million after 16 months commercial operation. These subscribers are great potential participants to SCM for dependable material flows: Electronic B2C. commerce is now in common; the next will be integra- In addition to the B2B and B2C growth as above tion of commerce and manufacturing and transportation mentioned, the scheduled future telecommunication of goods, i.e. inter-enterprise SCM (Supply Chain Man- platforms are viewed to bring the following three busi- agement). ness waves [7]. Physical processes including manufacturing and First wave: Digital TV, which implies penetration of transportation are vulnerable to malfunctions even TV commerce. However the scale will be limited be- though tremendous efforts attaining process reliabil- cause of rather small number of core players, i.e. broad- ity. In order to avoid damages caused by these mal- casting companies (around 100 companies in Japan). functions, material flows in market should be struc- tured as redundant so that some accidental services can be easily switched to available ones. Conven- tionally this switch has been done within each enter- Based on the observations, requirements for the fu- prise, but the revolution is giving opportunities shar- ture information systems are put together into three cat- ing workload each other in the market. egories as follows: In terms of information systems, this dynamic switch necessitates dynamic brokering capability of Service integration: Computerization of just one available services replacing malfunctioned services. business process is well matured. In stead, integration of many existing services corresponding to complex e-Government by one stop service: The developed business processes is strongly required. countries are now paying many efforts for computerizing government services in order to reduce government Dependable service delivery: These services are in expenditure as well as social cost. These efforts include the nature of publicity and therefore expected to be de- development of database for each agency, expansion of pendable. Dynamic brokering of services over the net- information delivery and processing services for legal work is one of the attractive means for attaining de- applications through the network, and arrangement of pendability in addition to traditional approaches for sys- legal systems for computerized administration. tems reliability.

Because of lagging computerization caused by Adaptability for knowledge creation: Systems that preparation time of social acceptance and less com- will be realized are in areas that we have never experi- petitive circumstances, presently each governmental enced. Therefore, we need to evolve in phases by learn- process is just on the way of computerization. How- ing and creating from experiences. Systems should be ever, in order to reduce the social cost essentially, flexible and suggestive enough for these learning and each administrative information system should be creating processes. collaborated through the network. For example, in case of passport application in Japan, presently people should take procedures to 3. Architecture of service integration many governmental agencies for getting residential platform certification and family register certification in addition to passport centers. If these procedures were 3.1. Basic concept handled in one stop shopping way, the social cost would be tremendously reduced. Based on the requirements summarized in Section 2, Service integration mechanisms for one stop shop- we put our basis for the information systems on the ping are definitely expected for e-Governments in global computing with knowledge creating steps. the present computerizing situations. The Internet can be seen a globally connected com- 2.3. Requirements for the coming information puting system with million servers, however the present systems use of the Internet is limited to remote job entry and re- We have been observing the coming revolution from mote database/meta-database access. Opportunity for bird view as well as actually planned information sys- collaborative services by million servers in the Internet tems that include both physical and logical world. Our should not be restricted to database access. Idea of ser- observations can be summarized as the follows; vice flow control with dynamic brokering is promising for integrating complex business processes. 1. The IT revolution involves every individuals and institutions, and provides opportunities to We have observed that knowledge creation for com- strengthen their activities. These activities cover munity and culture services is strongly required in fu- economy, community, and culture. ture. In order to respond to this requirement, a model 2. These activities should be dynamically woven for for social knowledge creation processes, SECI model social needs like conventional organizations such [9], is very useful, where SECI stands for Socialization as enterprises and governments have been work- (corporative working), Externalization (knowledge rep- ing. These activities also should be dynamically resentation of individual experiences of the working), woven for creating new values like conventional Combination (integration of represented individual social information structures have been function- knowledge), and Internalization (practice the integrated ing. knowledge). The SECI model tells that spiral work through the SECI processes is essential for knowledge assumed that the mediators advertise a set of at- creation. tribute names before the registration. 2. After the registration, the mediators make match- In the present situation, we can interpret the SECI es between the requesters and the providers ac- model as the follows; Externalization implies each ser- cording to distances between the request and service delivery and Combination implies service integra- vice profiles. tion, and these processes can be directly employed in our service systems. Socialization and Internalization The mediators initiate service delivery among the re- shall be conveyed by human, but can be supported by questers and the providers with customization according producing suggestive information as shown in later. to the requesters’ profiles, and monitor service delivery until the end of the delivery. 3.2. System model and architecture 3.3. Joint development of service integration In order to respond to the requirements, we have proposed ADSS model and architecture influenced by platform TINA business model. To realize a service system based on ADSS architec- ADSS model has three players with Ba that means ture, international joint effort for developing the service place in Japanese (Fig. 1). The ADSS model is intro- integration platform is initiated as detailed in Section 4 duced including requesters, mediators, and providers. of which code name is OSCAR (Open Service Collabo- The requesters are actors that correspond to consumers ration plAtfoRm). The mediator (OSCAR platform) in TINA, requesting and consuming services. The plays essential roles in our system, of which key func- providers are actors providing services as 3rd party ser- tions are illustrated in Fig. 2. vice providers. The mediators are actors providing ser- In the OSCAR platform, Planner, Organizer, and vices that mediate requesters and providers. In the other Flow manager are newly defined functionalities for words, the mediators have roles of retailers and brokers business process handling on the top of TINA model to in TINA [10]. The Mediation Ba is a federation of me- realize the target applications. diators where mediators provide mediation services to other mediators. In ADSS model, the service platform is considered Scenario as an environment that provides the functionalities for writers connectivity management and QoS management. In Planner Scenario this paper, we do not discuss the service platform itself. repository

Mediation Place (Ba) Organizer

Requester Mediator Flow Service Provider Broker manager providers End users Requester Core platform Provider (requesters) Mediator Requester Profile Other mediators Service Platforms OSCAR platform (mediator) Figure 1. ADSS basic Model Figure 2. Structure of proposed information platform Based on the definition of the players, we assume the following basic service provisioning and delivery Planner: translates service scenarios that are assort- process. ments of various service templates (classes of services) 1. The requesters register their request profiles with defined by scenario writers into plans. The Plans consist personal information as well as terminal capabili- of chains of service instances according to requesters’ ty to the mediators, and the providers register requests and providers’ service profiles. At this point, their service profiles to the mediators. These pro- we have introduced composite services as scenarios that files consist of attribute names and values. It is are delivered by the mediator. For the translation, goal programming or constraint satisfaction programming is thought one of the promising approaches for concise adequate service instance dynamically to guarantee a and reusable scenario description. service-level dependability.

Broker: finds out adequate executable service in- In these functions, the federated mediators are dealt stances (service offers) for the service templates in the with as the providers for the mediator. It should be also plans by profile matching, and corresponds to TINA remarked that the mediator might be wrapped for the re- Broker. questers and the providers for actual situations. This is because interactions between the mediator and the oth- Flow manager: interprets the Plans and manages ac- ers that are not equipped with expected interfaces are tual delivery of chained services by invoking respective occasionally required. service instances via a Core platform just like a work- In summary, the OSCAR platform makes it possible flow manager. The Flow manager also integrates ser- to integration of services distributed over the network. vices based on the Plans and delivers it to the Re- Its dynamic brokering capability contributes to en- questers. And it contains a set of Service Session Man- hancement of service quality as well as dependability agers (SSMs), each of them corresponding to a SSM in- and the function analyzing real usage of services pro- stance for individual interaction with Provider. vides baseline for knowledge creation.

Core platform: manages profiles and controls ac- 4. Collaborative development of proto- cess (sessions) to the mediator according to the sub- type scriptions by the requesters and the providers. In addition to these basic functions, it monitors service usage Japanese and German telecom and information in- and analyzes the usage for adaptation to real circum- dustries initiated collaborative development and experi- stances as baseline for knowledge creation and self-or- ments for the practical use are scheduled as the OSCAR ganization of services. The Core platform contains We- during this year. Participants’ specialties and main con- b-server and Adapters to inter-work with those systems tributions so far are as the follows; Hitachi has provided operating in Non-DPE environments. a composite service execution mechanism with experiences developing ADSS and the workflow. GMD

Chain of services Flight S Train E (brokering) Provider Requester Mediator Organizer Flight information and (mobile Mobile reservation system phone) carrier Provider Broker Flow manager Flight information and reservation system Provider Core platform Flight information and Requester Web -Profile manager reservation system Adapter (Web server -Community manager browser) -Access & log manager Provider Surface transportation CORBA information system

Figure 3. Prototype system based on the proposed structure for travel planning

Organizer: controls all other functions. In case that FOKUS has provided object oriented system architec- an fault of the service instance or the Provider occurs, tures with experiences on TINA and ADSS develop- the Organizer will call the Broker to select the second ment. NTT has provided a composite service description and translation mechanism for the Scenario and the Planner. NTT Data has provided knowledge on applica- 6. References tion and system evaluation with system integration experiences. [1] Ministry of Posts and Telecommunications, Japan, “Vi- sion of Information and Telecommunication in 21st Cen- tury -IT Japan for All-”, http://www.mpt.or.jp/, Mar. 2000 Application for the Experiments: The topic dealt (in Japanese) with is travel planning for flight and surface transporta- [2] PITAC (President’s Information Technology Advisory tion based on the Requesters’ and the Providers’ situa- Committee) Report to the President, “Information Tech- tion. Actually existing Providers are used including nology Research: Investing in Our Future”, http://www.- fright information services and a surface transportation cicc.gov/, Feb. 1999 information service that offers routes and schedules of [3] Electronics Technical Laboratory, Japan, “Ninf: A Global surface transportation upon the request with origin and Computing Infrastructure”, http://ninf.etl.go.jp/ destination. [4] H. Mulder (Ed.), “Business Model and Reference Points”, TINA-C (Telecommunication Information Net- Integration of carrier and surface transportation in- working Architecture Consortium), http://www.tinac.- formation services is done as shown in Figure 3, where com/, 1997 the Planner is not realized because the chain of services [5] ADSS-DSIG (Autonomous Decentralized Service Sys- is simple. By inputting a personal profile and a place to tems-Domain Special Interest Group, OMG), “ADSS-D- travel with expected departure/arrival time, the suggest- SIG Whitepaper”, OMG, http://www.omg.org/, 1997 ed flight and surface transportation are provided accord- [6] Hewlett-Packard Co., “Ten Ways to Think E-speak, ing to the dynamic brokering of the flight information http://www.e-speak.net/, 1999 services and the chained service retrieval. In this [7] J. Miyazu, “Towards Information Sharing Industry in demonstration, the integrated service can be accessed 21st Century”, presented at NTT International Sympo- sium, 1998 from mobile phones as illustrated in Fig. 3. [8] Y. Masuda, “Information Economy”, in Institute of So- We will develop the OSCAR platform prototype and cio-Information and Communication Studies, the Univer- its application jointly and will carry out an experiment sity of Tokyo (Ed.), “Socio-Information and Communica- to prove the usability of OSCAR platform. Then the tion Studies I System”, University of Tokyo Press, 1999 system architecture and interfaces of the OSCAR will (in Japanese) be opened and the international collaboration will be [9] I. Nonaka and H. Takeuchi, “The Knowledge-Creating called until the end of 2000. Company: How Japanese Companies Create the Dynam- ics of Innovation, Oxford University Press, 1995 [10] L.Strick, M.Funabashi, et al: “Efficient Brokerage in a 5. Concluding summary CORBA-based Service Mediation Platform”, Proc. of Expectations for the IT revolution such as three busi- ISADS99 - 3rd International Symposium on Autonomous Decentralized Systems, 1999 ness waves and three ECs are reviewed and discussed [11] Telecommunication Service Access and Subscription as participative society weaving individual efforts and RFP, (telecom 98-09-07) , OMG, http://www.omg.org/, creating new values. For realizing the expectations from 1998 the view of middleware technology, a service integration platform with dynamic brokering and flow control mechanisms is presented with a prototyped system that shows satisfying primal requirements. In addition to the technical discussions, international collaboration is proposed for enjoying the coming revolution. Finally, let us added that we are making international standardization efforts to realize our service system in actual use. We have formed an ADSS-DSIG (Autono- mous Decentralized Service Systems - Domain Special Interest Group) in OMG (Object Management Group) and issued RFP named TSAS (Tele-communication Ser- vice Access and Subscription) [11] that is a part of OS- CAR platform functionalities and interfaces for con- tract, access and subscription with mediator. TSAS is currently under specification mainly with the members of the TINA-C in Telecom Domain Task Force and will be approved in September 2000.