Summary

This, third, International Research Forum; to which Prof. Dr. Lutz Heuser, VP of SAP Research, invited a small group of participants to a facilitated debate on the future of an area of internet/business technology – this time “the internet of services”. It was held over two days in Potsdam near Berlin in late May and facilitated by Mark Jeffries and Dan Woods.

The presentations, a number of excellent photos, and a lot of background on the event is at:

http://www.international-research-forum.com/WIKI_2008/index.php/Main_Page

This note summarises my personal impression of the event which I found very stimulating and informative. The participants were both very knowledgeable and very forthcoming with ideas on how the “internet of services” might, or might not, evolve in the coming years.

This event is expected, like previous events, to lead to publication of a reference book to record the discussions, including follow-up research with additional experts.

My Conclusions? The internet of “services using web services” is coming from the cloud to our smartphones and carPCs soon. Some form of service directory including a trust measure will accelerate take-up which, in any case, will occur more quickly in service rather than manufacturing sectors. Cloud computing will be one of a number of “horizontal” supply chain elements that stimulate faster innovation of these information and entertainment services and facilitate a power shift to new service providers. Substantial areas of industry, especially those requiring high degrees of hardware, software and content dependability and stability, are likely to be left relatively unmoved by the phenomenon.

Introduction & Summary of 2007

Prof. Dr. Lutz Heuser (SAP) introduced this, the third, International Research Forum intended to look at some of the hype and trends around the “Internet of Services” using a series of paired vision and challenge speeches.

He reviewed some of the outcomes of the previous year’s event, not least the book, but also contributions to a EU FP7 project, the €90million Theseus “web-based service industry” project in Germany, and various events and conference input including CEBIT.

Introductory Session - The Web-based Service Industry

Dr. Neel Sundareesen (eBay) focused on “Services over the web using web services”.

Neel suggested that a successful service provider needed a broadband connection, inventory, shipping facilities, and a passion to sell.

He summarised some of the web service offerings that are changing things:

· storage, server, database infrastructure from Amazon, EMC, HP, IBM, Y!

· applications offered by eBay, IBM, Ning, Salesforce.com, etc

· platform offered by FaceBook, Google etc

· and soon - interfaces, user experience, social networks

These evolutions could make building a service as simple as authoring a webpage, although he believes that machine intelligence will always, especially in the long tail, need augmentation by humans.

He inferred that a future successful service provider will then only need a multi-function cellphone, inventory, shipping, and that passion to sell.

Dr. Stefan Wess (Bertelsmann/Empolis) countered by arguing that we already have all the necessary tools. Value creation is already moving into the long tail requiring knowledge-intensive business services which are growing twice as fast as commoditised services.

However, while we invest mostly to reduce the costs of processes, we end up spending more and more time fixing anomalies.

Using the example of arranging travel the 500km from his office to this conference he demonstrated some of the challenges in moving from today’s simplistic services to some really helpful services in the “long tail” (because they are lower volume / more personalised).

· Dynamics – reactive planning will be necessary to account for real-time disruptions etc;

· Richness – wide variety of service opportunities in the long tail; can we find them?

· Trust – how do we know which service providers we can trust?

· Accessibility – how will we discover new services? Google or a UDDI service directory?

· Tools – some exist but access is restricted by need for specialised knowledge

Stefan mapped a number of potential solutions against their potential impact on the challenges above and whether they were technology, economy, society or individual; arguing that the solution needed not only the internet of services but also integration with the internet of things, internet of people, and the internet of data – the semantic web.

The four breakout panels discussed issues as diverse as what are services?, the need for context information to be “clicked through” between sites, difficulty of modelling services - especially in the long tail, the role of government in encouraging take-up of web-based services. A consensus seemed present that there is a long way to go – some of the grammar exist but there are, as yet, few words.

Dinner

After dinner, Prof. Dr. Christoph Meinel gave a fascinating talk of the history and successes of the Hasso Platner Institut in Potsdam, founded by one of the two SAP founders.

Session – Infrastructure for Enterprise SOA 2.0

Dr. Werner Vogels (Amazon) suggested that infrastructure resources are being commoditised so that access to (massive) resources will no longer determine the success of a product and infrastructure resources will become variable rather than fixed costs.

• Basic storage, computation, and communications resources are being deployed by some companies, like Amazon, eBay, and Google, on such a scale that incremental costs are relatively low;
• These, and similar companies are able to offer massively scalable resources to users at prices that pre-empt resource ownership;
• Enterprise 2.0 companies will acquire and release resources on demand and only pay for what they use, allowing them to focus on building applications and providing services to end-user consumers and enterprises.

Amazon Web Services http://aws.amazon.com are an early instance of this phenomenon.

Dr. Matthias Kaiserswerth (IBM) challenged the vision with some concerns related to the quality and security of service and the ease of migration between infrastructure suppliers.

• Who gets favoured when there’s a problem?
• How difficult/costly is it to negotiate and enforce reasonable service level agreements?
• Will trusted intermediaries be required?
• Will this create a new market opportunity for the system consultancies?

Four breakout groups discussed the proposition, reporting back their summary findings:

• Owners of massive infrastructure will inevitably offer/sell spare capacity as a service;
• Expected service level will vary with price;
• Assisted migration between infrastructure suppliers could be a business opportunity;
• Users have to balance costs/risks of owning/leasing infrastructure;
• Questions around potential need for standards and regulation.

Second Session – Potential killer applications

Prof. Dr. August-Wilhelm Scheer (IDS Scheer) supported his contention that killer apps cannot be planned by referencing some killer apps that were intended for a different purpose:

• SAP R/3 based on AS/400 and aimed at SMEs, evolving to BigCO using Unix/SQL;
• eBay developed to exchange PEZ sweets, evolving to the standard for internet auctions

He also showed how specific applications often emerge slowly, building on something else:

• spreadsheets needed PCs;
• light bulbs needed electricity;
• iTunes needed MP3;
• DSL needed internet access;

He suggested several areas from which killer apps may emerge:

• Web2.0
• Cloud computing
• Semantic web
• RFID
• SOA

Prof. Dr. Martti Mantyla (HIIT) suggested that the killer app for the mobile Internet is already here in Microsoft Windows which facilitates the transfer of many Internet applications into the mobile Internet.

He seemed then to argue that this mobile internet platform that is always on, always with you, aware of where you are (GPS location), and your availability (calendar), and your social context (contacts list) can evolve as we add more functions (camera, audio, RFID ...), some connected by Bluetooth, so there is a gap between today’s reality and tomorrow’s potential that will create opportunities of new forms of social media and communications.

He explored some principles of social network growth from (personalised) things that attract your interest initially, through (user-generated content) things that hold your attention, to things that spread interest by exploiting your address book etc.

Four breakout panels took different approaches. Some focused on understanding what exactly is a killer app; others looked at what might make a Web2.0 killer app different; other sought out what made the tipping point of mass acceptance;

Some criteria included user-generated content, ease of use, rapid iteration (good enough code releases every day), mobility, context, location-awareness, personalisation, user benefit.

Third Session – Semantic Service Discovery

Prof. Dr. Wolfgang Wahlster (DFKI) outlined his views on the role to be played by semantic service discovery in making possible a Web3.0, as exemplified by work in the Theseus project.

He showed evidence that a hybrid form of semantic matching (in which a formally defined service requirement is automatically matched to one of a set of semantically-described services) generally outperformed matching based on logic alone.

However he commented/lamented that only around 2000 websites are currently semantically annotated so only a tiny fraction of the available services can be automatically discovered, let alone matched to a need.

Dr. Gautam Shroff (Tata Consultancy Services) thought there may be no problem as Google-type searches were more than adequate for most human-directed searches, even when only one or two search words were used.

He argued further that the use of text to define a search allowed the user to focus on the task in hand rather than worrying about the syntax of the query. This led to a case for natural language to be used as the interface to business applications.

Four breakout panels looked at aspects. Some picked up on the way in which services have to be defined in order to be automatically discoverable; others on the issue of trust – how do you know that an automatically discovered service is trustworthy or more dependable than an alternative; the value of social networks as a means of building a service provider’s reputation; how to distinguish a service trusted specifically by a small group of like-minded users – rather than the crowd. A conclusion was that the web could never be 100% semantically described. Some argued the case for a service directory, based on a combination of automatically and manually discovered services, supported by various measures of trustworthiness. Most agreed that there are currently no incentives to offer an independent guide to the trustworthiness of a service.

Crystal Ball

In the final session, participants were challenged to look five years ahead to identify how and where the internet of services might evolve.

Suggestions included:

• Widespread adoption of services delivered from the cloud with financial services being earlier than manufacturing;
• Web services in our cars provide information on congestion, speed limits etc;
• Personal Internet with our many aspects of our health tele-monitored in real time;
• No longer necessary to distinguish going to work and working from home;
• Internet’s key players will have changed;
• It industry will have responded to the challenges of climate change;
• Telepresence will have radically reduced the need for business travel;
• Internet technology will have to reduce its energy consumption;
• Internet of services will exist by then; or never will.

The event ended with a tour of the Sans Souci Royal Palace, a short tour of Potsdam, and dinner in central Potsdam.