On October the 10th, at 14:30 I’ll be chairing a roundtable in the Exchange Summit 2016 in Barcelona.
I am in the process of preparing this round table and learning on these Open Space type of events: as a chair I must drive the discussions and take notes on the conclusions from the experts, but there will be no formal agenda, therefore it is impossible to know what is going to be addressed during the session.
This will be challenging but I think that it will be better than a simple conference. We will be able to listen several opinions and learn from different initiatives on how these directories are being implemented worldwide.
I hope that relevant experts and interested parties on this topic will join our roundtable, and help having a fruitful discussions and conclusions.
This is an important topic because public and private directories for companies driving electronic business are a unique tool to discover trader partners and for establishing automated B2B connections.
The best way to boost the adoption of B2B is providing directories: that will allow users to find new opportunities to enlarge their network of B2B connected trader partners. The old-fashioned way of signing connectivity agreements from the EDI Value Added Networks world seems to start fading out, and the new and agile way of multi-partner agreements is step by step taking off.
An example of this multi-partner agreements is the PEPPOL network, where the TIA (Transport Infrastructure Agreement) sets the rules for the actors participating in the B2B exchanges within the PEPPOL community. Once you join the PEPPOL community, you know there are established semantic models, processes, syntax implementation and a transport mechanism that you need to follow. The rules cannot be broken, and the counterpart is that you are then allowed to drive electronic business with everybody else already registered in the network.
One of the important characteristics on the PEPPOL architecture is the distribution of the registry of participants. Even if there is a central SML (Service Metadata Locator) where you can query for a participant identifier, this DNS-like service only points you to the SMP (Service Metadata Publisher) that holds the company metadata information where you retrieve information about the reception capabilities and the endpoint of the receiver. With this architecture, it is difficult to search a particular participant in the whole network if you do not know the identifier he has used to register in PEPPOL. You can’t just say “I want to know whether my supplier Invinet can drive electronic business”. To answer this question, you need to know the identifier Invinet has registered in PEPPOL (you can use different identifier schemes in the PEPPOL network), you need to query the SML, and then the SMP in order to know which documents Invinet is able to process automatically through the network. Hence you need to talk with Invinet in order to discover his participant identifier.
In order to solve this issue Ger Clancy from IBM started an initiative called PEPPOL Directory. The scope of the PEPPOL Directory project is to enlarge the data for the SMP in order to provide more information on the participants, and to create a central point to collect this information, allowing for the global search of trading partners. But there are some important topics for discussion in this project, and even if I know that the project is still on-going, I’d like to start raising just three issues or concerns that are on top of my head:
- Should we build a Directory as a centralized database ? Centralization is always the easy way to do things, we have seen this approach in many projects, but I think it has more drawbacks than benefits. Decentralization provides more robustness, security, distributed management and responsibility and increases flexibility. The PEPPOL network of SMPs is an example of distributed network, unfortunately, PEPPOL has also a centralized SML service — I remember discussing this topic with Mikkel Hippe Brun back in 2008 in the first CEN BII workshop in Stockholm, when we envisaged a DNS-like type of transport architecture, at that point my idea was more in line with the current distributed Internet DNS than with the private centralized DNS it has finally been implemented –. Another issue with the centralized database is the scope: Should this database be national? European? International? Every step up makes the project more complex. And there is also the ownership issue… who would own the database?
- What metadata should be in these directories? This topic has different implications, for instance, in the legal domain, there are different data-protection laws in different States and they should be taken into account. From a pure maintenance point of view there are also a lot of challenges depending on the number of data in the directories: who and how will maintain the information so it is always up-to-date. The more information in the database, the more complexity to maintain it.
- Which use cases should be covered by the Directory? Is is only about searching companies? Should we be able to know which business processes the parties can support? Should the Directory evolve in something like a Global Company Registry? In the PEPPOL network, the SMPs only registers the participant identifier and the reception capabilities. They lack the sending capabilities, and then we lack use cases such as “discover which of my trading partners are able to send me invoices”, in order to ask them to start sending electronic invoices with my company.
And these questions are just for the PEPPOL Directory. There are other initiatives implementing public and private directories. I do not know them all, but for instance, in Spain the Directory of public entities is called FACe, and in Norway there is a particular PEPPOL SMP called ELMA where public entities shall register their capabilities. ELMA works similar to a general SMP, but FACe is using a very different approach, where the identifier for the participants is not a unique number but a set of three different identifiers called DIR3, and there is no registry of capabilities as this is only intended for electronic invoice submission to the public sector.
These topics or others (you know Open Space!) related with public and private directories will be discussed in Barcelona next 10th of October. I hope we will engage people from different directory projects explaining their point of view and help solving with our conclusions challenges such as the global reach of the e-business directories and the use cases they should be built for.