This is a short overview of the privacy issues for customers in the Smart Grid infrastructure. Smart meter and other devices within the Smart Grid produce a lot of privacy sensitive data. With this nd grained data it is possible to make predictions about the daily routine of a household or create a movement pro le of a vehicle. There are techniques to protect the privacy of a person in network-alike structures, e.g. by creating pseudonyms or anonymizing the ow of data. But this protection does not always work in an adequate way, for example if there are quasi identi er, signi cant or linked pattern, it is possible to depseudonymize and de-anonymize a customer. So it is possible to analyse the daily routine of a person as well as creating a movement pro le of his electrical vehicle or getting some informations about his preferences or issues.
A consumer, can also appear as an energy producer, as Prosumer. To stabilize the grid and protect it from overload and under-supply, it is necessary to keep the consumption and production of energy in an equilibrium. Therefor the Energy Service Providers (ESPs) and the Smart Meter Gateways (SMGs) of a Prosumer are directly connected over the Internet. As a result the acquisition of sensor data to the split second of consumed and produced energy is possible. Also controlling data can be submitted from the ESP to the Prosumer. So the Smart Grid initiate a paradigm change from the pure power grid to a combined power and communication grid.
This results in the following requirements for the security and privacy which
are prescribed by the German BSI [
In the future the Smart Grid will be a signi cant part of our life and everyone
have to participate and interact with this construct. The industry wants to collect
detailed, ne grained meter data of customers consumption. But each way of
life and behaviour is individual. Thus a person or a household share a lot of
information of its way of life with his ne granulated energy consumption trace.
Figure 1 from Newborough and Augood [
One additional component of the Smart Grid infrastructure will be electric
vehicles because they can be integrated into the infrastructure in a useful way.
The idea of the Vehicle-to-Grid (V2G) concept is to use the batteries of the
vehicles as centrally coordinated, distributed grid resource [
The privacy problem would be solved, if every consumer has the same and steady behaviour and consumption, so no individual behaviour could be identied. But this is not a realistic postulation and we have to nd technical solutions, which protect the privacy on the one hand and provide the necessary data for the industry, e.g. for billing, on the other hand.
To protect the privacy of the Prosumer, the BSI claims that anonymisation
and pseudonymisation techniques must be used [
But the protection of the anonymity over a long period of time is not a trivial
request and this problem is not solved yet. Because a anonymized connection
does not always protect the privacy of an user. Other research work has
demonstrated, that the longterm aggregation of partial information from anonymized
users can break the anonymity and reconstruct the user pro le [
Another privacy protecting method is the creation of multiple pseudonyms
of a Prosumer. The pro t for the Prosumer would be that, if one pseudonym
is uncovered, only a small part of the Prosumer data could be assigned to him.
But multiple pseudonyms has also aws. Jawurek [
The linking by behaviour anomaly attack can be used to link either an identity to a consumption trace or more then one consumption traces together. An anomaly de nes a singular or a series of unusual events. If this anomaly is reected in the energy consumption and individual linked to the customers behaviour, an identi cation is possible. As an example the leaving and coming home times could be such a linking parameter. This events need to be collected in a high-resolution. But also a low-resolution identifying event could be used to identify a user. For example, if the inhabitants leave every weekend or stay at home on speci c work days.
The linking by behaviour pattern can be used to link di erent pseudonyms of one person. A customer can have multiple pseudonyms. For example a new pseudonym is generated, if the supplier is changed. One other situation could be, that the supplier wants to protect the anonymity of his customer and generates a pseudonym after a certain time. For example the pseudonym A is used for the time interval t and the pseudonym B for the time interval t + 1. The bene t of this would be, that if an identity is de-pseudonymised, only a nite period of time is compromise. An attacker could now try to nd a signi cant pattern inside the consumption traces. If this patterns are found in other consumption traces of the customers pseudonymised identities, the pseudonyms can be linked. 3
The Smart Grid brings an innovative change for the electricity market and his
participants. But, as discussed, this change go along with risks for the privacy
of every customer. It would be desirable that the Prosumers have the ability to
protect his own privacy but for changing his individual behaviour. So the solution
should be nd on the modality, how the consumption data are processed and
used. Jawurek et al. [
The next step will be the analysis of the question, how detailed and ne
granulated consumption data must be collected. One part of this step will be the
investigation of Intrusion Detection Systems (IDSs). IDS are based on the
analysis of ( ne granulated) information ows and IDSs are necessary to protect the
Smart Grid and the Prosumer, because there are a lot of threats [
A de nition of an IDS is given by [
One possible method could be the anomaly detection which was early
described by Denning [