DLMS USER ASSOCIATION Why a single data model is key to the success of smart metering

Position paper 26th April 2012

Introduction
The IEC 62056 DLMS/COSEM series of standards developed by IEC TC 13 and supported by the DLMS User Association provides a consistent standardization framework for smart metering. The gist of this framework is to use a single, communication media independent, object-oriented data model that can be used over a wide range of communication media, and allows achieving interoperability on the end device level. In the smart metering / smart grid context, standardization became essential to remove market barriers and to facilitate large scale rollout at lowest cost. Consequently, some technologies that have been kept proprietary before are proposed for standardization. Those technologies come with their own data model creating a conflict with the established standardization framework and jeopardize the consistency of standardization. For some other technologies proposed, only the communication media dependent lower layers are specified. These technologies can be readily integrated into the IEC 62056 framework. This situation offers two options for handling new standardization proposals, as illustrated in the Figure below. The purpose of this paper is to provide arguments supporting the IEC 62056 framework and for consistent standardization.

Right hand side of the figure above: Each new communication technology is integrated into the IEC 62056 framework, using ONE standard data model and ONE set of data security tools. This is a consistent approach in that: The data model (use case dependent) and the communication protocols (technology dependent) are separated; The object oriented data model specifies how the various use cases can be supported and how the model can be extended to support new use cases; The lower layers provide maximum flexibility to support a range of existing and upcoming communication media.

Left hand side of the figure above: Some standards proposals are Self-contained specifying both the lower layers and data models. These proposals are in conflict with the existing standard and also with each other. Their promoters argue that interoperability could be achieved on higher levels of the system, so inconsistency on the device level is a non-issue. In the following, the different aspects and possible consequences are explored. Standards must be consistent; conflicting standards are against the very purpose of standardization On the one hand: Standards have to meet certain requirements: The purpose of standardization is to bring stakeholders together to agree in an open and transparent process on solutions that will be generally followed. Standardization organizations must make sure that what is called a standard is the result of consensus based standardization processes the organization stands for; It is the task of the standardization organization to provide universal, consistent standards and not a set of conflicting standards; Participation in standardization is an investment from all stakeholders, which is returned if the standard creates fair market conditions and level playing field for all. This investment must be protected; Participation of many stakeholders forces writing precise specifications which everybody can understand and validate; Standards resulting from good processes can be implemented by anybody with the right skills, and this creates a large supplier base; Widely accepted standards also create a market of support tools and services, like software libraries, test software, training packages that facilitate the smooth implementation and application of the standard.

On the other hand, there is no evidence that home grown specifications meet the requirements for a standard: The platform for competition is the market, not the standardization organization. It is left to the market participants to invest into a universal standard solution or into a manufacturer specific solution both having its advantages and disadvantages; A standard is not the publication of a manufacturer specific solution. A serious standardization organization is not a publishing agency for manufacturer specific solutions. Publishing such specifications, developed in an intransparent process would be against the very purpose of standardization; Technical committees approached with such proposals that ignore the good practices of standards development are faced with the impossible task of verifying the claims on openness, completeness and correctness of the specification. Without these properties, implementation by many parties is not possible; Promoters of home-grown solutions may claim large installed base as an argument for accepting their solutions. However, standard committees have no means to verify that the specifications delivered are consistent with the installations referenced; Publishing manufacturer-specific solutions as conflicting standards would destroy the value of the standard label; Should a Technical Committee embark in accepting such proposals, this would invite more of those kinds of proposals. This would lead to a fragmented market. Utilities having invested into a standard solution would lose their suppliers because now every supplier will only support its own standard.

IEC 62056 DLMS/COSEM has been developed in an open standardization process, It is the result of huge investments by many stakeholders including manufacturers and purchasers. IEC 62056 has indeed created a level playing field. For this reason it has gained wide acceptance: Today (April 2012) the DLMS UA has some 250 members, the standard is implemented by some 70 manufacturers and there are some 230 compliant implementations. There are support tools and services available from independent sources. Why alternative data models are not necessary A data model describes what data have to be present and exchanged to support the various use cases. Examples for use cases are managing customers contracts, reading of billing data, configuring tariff schedules, controlling the supply, power quality monitoring, event handling, providing information to the consumer, firmware upgrade just to name a few. Use cases are identical or very similar on each market. Modeling the same use case in different ways does not add value but hinders interoperability; The IEC 62056 COSEM data model has standard extension mechanisms to support new requirements. Since the first publication of the standard, many new elements have been successfully added to support smart metering and smart grid applications while keeping full backwards compatibility;

Should further market requirements or developments in technology be not sufficiently considered by the existing standards, the standardization organization has the right processes in place to improve and extend them in a consistent way; COSEM is an established, integrated part of the IEC smart grid architecture: o COSEM (IEC 62056-61/62): data model for revenue metering equipment; o IEC 61850: data model for grid automation equipment on substation level; o CIM (IEC 61970 / 61968):data model for data exchange between the utilities ERP system componentsThe recently established TC 57 / TC 13 joint working group JWG 16 is specifying the link between CIM and COSEM; Proliferation of data models for revenue metering would necessitate multiplying the efforts for mapping these models to models of related domains.

Efficiency of standard and optimized solutions On the one hand: Universal standards have to be designed in such a way that they can satisfy a wide variety of differing market needs. Therefore, they should prescribe only a few mandatory elements, allow suitable options and must provide standard extension mechanism to cater for future needs. This certainly leads to some overhead. However, the overhead can be minimized if project participants agree on the mandatory elements and options needed. Such agreements may be fixed in companion specifications (see later).

On the other hand: Self-contained and/or Home-grown solutions are often optimized for a given target application, therefore may be very efficient. However, as soon as such specifications have to be used in a different context for another application, it is difficult or impossible to accommodate new requirements, or the efficiency is lost.

IEC 62056 DLMS/COSEM provides many ways to ensure efficiency despite of being universal and flexible. Achieving interoperability: the importance of conformance testing To prove that an implementation meets the requirement of a standard, test methods have to be specified and test tools have to be available; As the Home-grown solutions are not meant for implementation by many different vendors, test methods and test tools are generally not available.

The DLMS UA has specified conformance test plans for the various parts of the IEC 62056 series and runs a conformance testing scheme. To date, some 69 companies have DLMS/COSEM conformance test laboratories. 13 of these are performing third party testing. Some 70 manufactures have obtained Conformance Certificates and some 230 devices have been approved. The list of DLMS/COSEM compliant equipment can be found at: [Link]/conformance/listofcompliantequipment/[Link]

Achieving interoperability: The role of companion specifications As already mentioned, a standard should specify only a few mandatory elements but several options to cover the widest range of applications, and to allow extensions to cater for future market needs. There may be interoperability gaps between implementations that have taken a different set of mandatory elements and options from the standard; To close such gaps, the standard can be customized for a given application / project purposes. These customizations are the scope of companion specifications; It is the interest of the buyer to involve several suppliers in developing the companion specification and to make it available for potential suppliers; Implementations of companion specifications are fully in line with the base standards, and ensure full interoperability and optimized performance.

There are many examples of companion specifications to IEC 62056 DLMS/COSEM, on which smart metering projects are based: e.g. Linky (ERDF / France), DSMR (Netherlands), PRIME (Iberdrola, Hidrocantbrico-EDP, Gas Natural Fenosa / Spain, Hidrocantbrico-EDP / Portugal), IDIS (covering several markets), in Saudi Arabia and in India. Can interoperability also be achieved with different data models? On the one hand: End devices may be all different, interoperability can be also achieved at higher levels of the system. This is the typical argument of a meter manufacturer promoting its manufacturer-specific solution. And it is certainly true, an entire industry the system integrators base their business model on this.

On the other hand: Once a utility buys in it is locked to this meter manufacturer; i.e. meters from other manufacturers can only be integrated into the system with substantial integration costs. This also decreases the number of potential suppliers therefore reduces competition; The purpose of achieving interoperability on device level is simply to avoid the high costs of system integration and to keep competition open.

IEC 62056 DLMS/COSEM brings interoperability with this, manufacturer independence on the end device level, the highest volume system component. Building systems with interoperable end devices can substantially reduce costs and risks for the utility. Summary The DLMS UA believes that using a single, standardized data model for revenue metering is essential for the successful rollout of smart metering systems. It is the key to achieving interoperability at device level, to avoid high system integration costs and fair competition. The DLMS UA, in co-operation with the IEC and other bodies, is committed to keep the standard up-to-date to meet new requirements and be in line with the latest technology developments. It provides an open, transparent platform for its members participating in the development to ensure the high quality and consistency of the standard. It is also committed to provide a high level support to the users of IEC 62056 DLMS/COSEM.