Electricité de France (EDF) produces and sells electricity all over France. This activity may readily be defined as ‘service providing’ instead of ‘selling goods’ – in fact ‘energy’, because electricity is a very special ‘good’.  It cannot be stocked by customers (unlike other fuels providing energy), it cannot be delivered in advance, and one of the most important aspect of the job of electricity utilities is to produce and deliver at any instant the exact amount of electricity that the customers need. An electricity utility must thus manage securely, in real time, all its production resources (power plants and distribution grids) without the customer even being aware of it.

More recently, the awareness of sustainable development encouraged EDF to initiate R&D activities about correlated services that could help consumers to get more from their electricity (to use it in better ways) and to adopt behaviours that might be globally profitable, with a view to reducing the societal costs of producing and delivering electricity. This in turn would allow electricity utilities to manage their global resources more efficiently, and marshal their development more economically.

This was our main macroeconomic motivation for research on smart distributed energy management techniques, which will actually be conditioned by progress in the more generic smart home services domain.

TEAHA (The European Application Home Alliance) (www.teaha.org) is a European project (a STREP) that addresses networked home control applications and how they complement A/V networked applications. Its objective is to work with the A/V world to specify an open, secure, interoperable and seamless global home platform. The TEAHA initiative has been structured into two phases. Phase 2 will involve a significant test programme.

During Phase 1, the route to industrial deployment will be addressed through the validation of three applications – white goods control, energy management, and residential gateway applications with seamless integration with networked A/V applications. Thereafter standardisation work will be carried out, and finally marketing and dissemination will involve preparing TEAHA Phase 2.

Metering services for smart home appliances

Metering services for smart home appliances

It must be noted that, to ensure seamless integration of services, one important commitment of this project is the adoption at this level of the technology standard of OSGi (see box).


To take the discussion further, one has to make some generalisations about the business relations between the electricity utilities’ tariff policies and local energy and cost management. In a few years, there may be at least two kinds of actors in this market.

  1. Energy producers (including the power plants and the distribution grids).
  2. Energy sellers (who will manage the business relations with the customers).

And of course there will also be the customers.
The interactions between these actors will tend to be complex, but we could start with the following simple sketch:

The way customers use electricity (how much, when and where) will result in production costs for the producers, related to the efficiency of the power plants they have to keep running at a given time, and to the safety of the exploitation of the distribution grids. When utilities base their actions on predictions about customer behaviour, the long-term result is likely to give rise to increased investment costs.

Both these costs could be minimised if customers use more or less the same amount of energy all the time (this would allow the most efficient power plants to be kept running steadily) and if the electricity is produced not far from where it is consumed (so that long range energy flows could be limited).

Adopting the OSGi standard fosters collaboration

Adopting the OSGi standard fosters collaboration

Energy sellers can be viewed as ‘syndicates’, allowing their customers to negotiate best conditions with the producers, or as traders who buy the electricity at the best price and can thus offer their customers the cheapest tariff. But in any given geographical region there may be several competing sellers, and consumers may change supplier if they find a better offer elsewhere.

B2B relations between service providers

B2B relations between service providers

So energy sellers have to interact with two different groups – the producers and the customers. They will be able to negotiate the best conditions with producers if they can prove that they ‘own’ customers whose global behaviour is likely to help the producers reduce their costs (and investments). To attract consumers, they will have to develop tariff offers with incentives that yield the desired global behaviour, and persuade them to adopt the related consumption constraints and become their customers.

Such incentives would be the most objective way for energy sellers to compete, both in their negotiations with the producers and in their efforts to gain the biggest share of customers, and thus develop the biggest business. One can foresee that these incentives will vary widely, as has happened with the diversity of the TelCo or ISP offerings.

However, we know that advantageous but complex tariffs cannot be sold when the customers have to manage the resulting constraints without the help of automation tools. Customers would have to check the time continuously, as well as identify all the appliances that are consuming power at that time. If after a while a customer notices that he does not really save money, or that the constraints are too difficult to manage, he will change to a simpler tariff offered by a competing seller.

In order to simplify the management of complex tariffs, customers can invest in a smart home software system with a lot of functions – among them a curtailing component that is exactly suited to the tariff constraints. During peak hours, this component receives over-consumption alarms from the tariff checking component and may decide to automatically shut down some of the appliances it is allowed to. Of course, the over-consumption alarms are generated in real time, and take into account the unpredictable actions of customers on any other appliances. In addition the curtailing algorithm will take these decisions ‘smartly’, because the smart home component also communicates with appliance managers, which can tell it “Sorry, you cannot interrupt the washing machine now – please stop another appliance.”

To sum up, energy sellers will propose various tariff saving offers, ranging from simple bi-tarification to real-time energy trading (eventually even with progressive rates depending on instant consumption – why not?) They will do so in an effort to encourage customers to adopt the behaviours that allow the sellers to negotiate from a position of strength. But to get their customers to accept these offers, they will also provide adapted saving strategy tools that plug into smart home software systems. So the liaison with the home automation service provider community is the most important issue.
The three conditions of smarter energy management are:

  1. That the utility (or seller) provides the needed real-time tariff signals.
  2. That the smart house provides sophisticated saving strategies.
  3. And that it has real-time access to the consumption data.

OSGi glossary

OSGi* (Open Service Gateway initiative) is an alliance of companies dedicated to the elaboration of a standard (the OSGi standard) defining technical means to develop, to deploy, to administer and to maintain network based residential remote services. The OSGi standard is focused on the services, i.e. on the application layer. It claims to stay neutral toward all the underlying technologies.

Framework : Executive environment of an OSGi gateway.
An OSGi gateway is any Java machine with suitable hardware connectivity to be linked to home networks and to external networks of any suitable kind in which an OSGi framework is running. The main task of this framework is to manage the life cycle of software components installed by service providers in the gateway, and to provide the seamless integration and co-operation of the deployed services.

Bundle : The unit of installation of services in a gateway.
It is on this scale of bundles that the service providers negotiate the installation of the software components they need in the gateway with the OSGi framework. The bundles can comprise one or several Java packages and all kinds of other content/resources like images, sounds, HTML files, applets, data bases, etc. that the offered services need to have local (within the gateway) access to. It has to have a manifest file that describes its content and properties, and an activator class that implements the specific ways to start and stop the software components that were installed.

Activator : A class attached to a bundle for activating its contents.
The activator class has two methods : start( ) and stop( ). The start( ) method is called by the framework when it has installed the bundle and succeeded in resolving all its dependencies. It will register all exported services of the bundle, instantiate references to all the registered services that the bundle may need and instantiate and activate any objects of its implementation packages. The stop( ) method does the reverse, as needed for cleaning up.

Service : by the OSGi meaning exported interface that may be registered by the framework.
In this way, the (abstract) methods of a service and its (static final) attributes are made accessible to other software components installed in the gateway. This ought to permit the seamless co-operation between bundles from various service providers – for instance, metering services installed by the power meter owner could be used by energy (and cost) management software, or real time consumption models identification. Due to the OSGi relevance, services which are interfaces or specifications are open or ‘public’, and thus accessible by all, whereas their implementations are ‘private’, which means they are the property and the business stakes of the service suppliers.

* Note: the OSGi standard has also gained other domains of application and use, for instance automotive electronics (where its acceptance is even more dynamic!).


At this time, the EDF/EURIDIS electronic power meters that are already deployed in millions of French homes support the third condition rather well.

These power meters have two ports: one of them – if properly configured – sends metering information continuously (at a sampling rate shorter than two seconds) and may be linked to customer devices for smart home management.

Power suppliers are most interested in using this potential to offer innovative tariff management that is truly versatile and proactive. But the valuable data output from the power meter could be sold to other service providers, to implement smart adaptative energy/cost/in-house-load-control tools that could, moreover, be designed to co-operate if needed with the billing services to implement customer-managed cost saving strategies.

Based on these hardware specifications, EDF has already developed basic metering software components in the form of OSGi bundles (to be downloaded into an OSGi residential gateway) which consider these basic metering facilities as OSGi services. There is one bundle, called the metering services bundle, that abstracts these ‘services’ from any hardware and communication link.


The EDF/EURIDIS electronic power meters may be viewed as very useful ‘appliances’ in a smart home. They can provide all kinds of consumption parameters with a fast sampling rate, compared with the timing of most domestic processes.

But because the energy management system of a smart home has to be global, it has to process information coming from other appliances (in fact all appliances it can ‘speak to’) to make arbitration decisions for real cost saving without disturbing any of the people living in the house. These people will probably not even be aware of which automation tasks are running at a given time, or of what they do. The decisions taken by the energy management system may even change when the customer chooses new tariff options with new constraints, or when he buys new appliances that will let the smart home manage the constraints in other ways.

So the main issue is to specify the minimal and generic ‘plug’ on the smart home software system that can accommodate various ‘plug-in’ components, paired with any tariff/pricing policy that can be imagined by sellers and adopted by different customers. And on the other side of the smart home system, it is necessary to specify the ‘part of the plug’ related to smart energy management (with conditional curtailing, but also other possible functions) to which one could attach specific ‘device management plug-ins’ that care principally about the appliances to be managed safely, and thus do not allow an appliance to be disconnected when safety or optimum performance do not allow it.