Rodrigo del Busto, Senior
Analyst Model Development,
Origin [HREF1] , Lv. 45 Australia Square 264-278 George st. Sydney NSW 2000 Australia.
rodrigo.delbusto@originenergy.com.au
Andrew J. Bucknell, Faculty of Engineering, University of Technology, [HREF2] Sydney, P.O. Box 123, Broadway NSW 2007 Australia , Andrew.Bucknell@uts.edu.au
Energy Traders need to understand, and react to, large amounts of information in order to do their job effectively. The nature and variety of these information sources means that traders must spend most of their time manually sifting through data instead of planning strategies and making crucial decisions in a 24 hour market that is extremely volatile.
By using WEB 2.0 technologies, this project used a SaaS (Software as a Service) approach to cost effectively build a series of small applications in-house, which delivers aggregated information in a simplified format to the Trading community in a more cost effective and flexible way than using off the shelve software.
Typical approaches require the deployment of expensive/complex application frameworks. Our approach shows how to achieve similar results on time and at lower costs by using dynamic html and iframes to present information to the user. The utility of our approach is demonstrated through a case study showing how we integrated diverse unsynchronized data sources in to a near real-time information system that supports rapid decision making
To satisfy the demands of energy traders for dynamic information requires flexible responsive systems. The systems must not only be flexible and responsive with respect to the way they display information, but they must also display these characteristics of the information sources they handle. We have developed a Web 2.0 portal that addresses this problem in our environment. A key feature of the approach and infrastructure used to implement our portal is that we can readily develop new components that make new information sources available.
The portal was developed in response to pressing information needs within our organisation. The work here describes how we developed the portal, the lessons learned from this development, and the implications of this development for future research and development, both within our environment and within the wider community. These points are presented in the context of a case study presenting the Trading Risk Analysis Portal (TRAP) system we have developed and deployed.
Our System brings together a number of information sources and presents them to our Energy Traders to help them assess and manage market risks as they develop. The problem we identified in our organisation was the time required for traders to manually access a diverse range of systems to access the information required to make a decision. As shown in Figure 1, the time required to access and synthesise this information was 15 minutes. In an environment where trade positions are finalised every 30 minutes, this time was too long. Time lost navigating between different sources each containing large amounts of information, increases the potential financial exposure of the Traders position. Our goal was to automate the access and synthesis of information in to a single application, as shown in figure 2. By doing so we significantly reduce the amount of time traders spend seeking information, allowing them to react more quickly to changes in the markets. In order to achieve results we based our approach on that taken by Netvibes, [HREF4] or iGoogle [HREF5] of building small widget like applications and presenting through a series of portal pages. A portal-based approach has been identified as being an effective strategy for integrating diverse information streams in to a single application (Wege 2002).
Figure 1 Trader Information Seeking |
Figure 2 Trader Information Seeking with TRAP |
We achieve near real-time response by using Software as a Service (SaaS) approach for data collection and redistribution where "…services are assembled and provided as and when needed to address a particular requirement" (Curbera 2001). On a technical front, this approach is appropriate as it facilitates ease of maintenance and extensibility. From a business perspective each application aggregates information relating to a potential source of risk and presents it in a simplified format to Traders
To address the shortcomings in our existing system we developed a web system to process the vast amounts of information available. We identified three key requirements for our system. Our first requirement was that users had to be able to access the system without the need to install client software or update as new versions became available. The second requirement was that new data sources and visualisation can be added to the system as they became required or available. Lastly, the system has to be easy to manage within a secure framework.
The first requirement was addressed by using a web-based user interface. This means that users can access our system by using any web browser, and as a result access to the system is not restricted to particular machines or configurations. The second requirement was addressed by having a number of portal-like pages, each presenting SaaS based views onto the information sources. The advantage of this mix of client side and server side technologies is that this approach leads to pages that are easily maintained and extended and customised to different screen resolutions 1200x1024 screen on a normal desktop or 1024x768 for the large LCD screens.
The final requirement was addressed by leveraging of existing infrastructure provided by ISS and windows domains. We used integrated authentication, as the security provided by it was sufficient, as the application will only be accessible within the company's intranet.
The requirements for our solution are known to be continuously evolving, based on our experience providing applications to the traders. Any solution we deployed needed to be able to meet not only our existing requirements, but be able to be adapted to new requirements. We therefore decided early our solution should be based on a software framework, as such an approach would give us the reusability and extensibility we required (Fayad 1999). When evaluating suitable frameworks for the deployment of our system we where not able to find a ready available solutions that satisfied our key requirements of accessibility, expandability and security. Publicly hosted solution, like those of Netvibes and IGoogle, were considered as they offered the accessibility and expandability required, however they failed to adequately provide a security framework for sensitive information.
The sensitivity of the information was the most important requirement, and it imposed in a number of restrictions on our solution architecture. Firstly, we concluded that the application had to be restricted to use within our intranet and remote VPN. Second, we were restricted by the company's Standard Operating Environment (SOE) in the technologies we could use. Finally, tight controls had to be in place to restrict access to different user groups to each Widget application as well as to the portal pages. An effective security model was implemented by leveraging our organisations existing integrated authentication provided by ISS and windows domains.
Once the security framework was in place, we had to decide on how to effectively implement accessibility and expandability. Expandability encompassed the principles of reusability and separation of concern, which have been identified as being important to web-based frameworks (Schwabe 2001). In order to integrate the diverse technologies into a single system we imposed three key constraints on the design of components. First, each application was treated as a stand-alone DHTML unit, with no dependencies on other DHTML components in the system. Second; look and feel was governed by style sheets, enforcing a separation of user interface from the data being presented. Finally, we enforced a rule that each component manages its own database interaction. While this approach prohibited reusability of data communication layers, it meant that the data layer of any component could be modified without having to worry about potential impacts on other components in the system. These design constraints allow us to develop and maintain components independently of one another, meaning that we can readily add new or modify existing components.
To develop components based on these design principles we applied a Rapid Application Development (RAD) methodology. Users were extremely receptive to this approach because we had a fast response time to implement their requirements and respond to any requests for changes. These approaches to developing technology satisfied the requirements we had identified as essential for successfully publishing information within our organisation. Beyond the technical issues, a significant contributor to the success of our approach was the conscious identification and cultivation of community within the stakeholder groups. It is through this combination of technology and community engagement that we have been able to develop our framework successfully.
The framework-based application we have developed provides information to Origin Energy traders. These traders require up to the minute information so they can make decisions that pre-empt volatility in the market. In particular, real-time information helps traders react to occasional price spikes by identifying patterns associated with such spikes and changing their position accordingly. To assist traders with this decision-making we have built and deployed the TRAP information system. TRAP integrates diverse information streams into a single portal. A screenshot of the application, showing a range of applications grouped onto one page, is shown in Figure 3. The implementation of the TRAP application has been based on the system and framework principles described above, and demonstrates the utility of these concepts.
We have used Web 2.0 technologies to integrate these information sources into a portal-based application. Applying the principles described above we have been able to tailor our set of widgets to suit the needs of our traders. The approach we adopted was heavily influenced by the NetVibes Universal Widget API (UWA). By creating vertically independent components, we have been able to tailor the information displayed to the groups of users who use TRAP. By engaging with the stakeholders and using RAD methods, we are able to work with the stakeholders to meet changing information needs. Our framework makes use of community involvement and a set of development guidelines to allow effective ongoing growth.
Because of constraints on data security it was not suitable for us to use publicly managed portals such as iGoogle or NetVibes. Creating a framework in-house meant that we are able to maintain ownership and security of our data while taking advantage of the extensibility and flexibility of a portal-based framework. We found we were able to easily achieve the functionality of such a framework by applying readily available technologies such as DHTML, and client- and server-side scripting. The use of script-based technologies meant that we were able to be to be responsive to changes in user requirements and were able to make changes to the system while discussing requirements with stakeholders. This flexibility and responsiveness was a key factor in the success of our implementation.
Electricity trading markets are extremely volatile, and in order to handle this volatility in a profitable way traders require near real time information presented in a way they can easily digest and use to make decisions. An effective strategy for presenting this information to traders is through a web-based portal that aggregates diverse information sources in to a single information source. We needed a framework that would support this aggregation in a flexible, secure and extensible way. No existing frameworks were found that met all of these requirements, and so we constructed our own. We found that using Web 2.0 technologies allowed us to meet these goals, and also encouraged collaboration fostering a sense of community involvement with the application. This combination of technology and community resulted in a successful implementation and deployment.
Blanco C. and Soronow D. (2001). 'Jump Diffusion Processes - Energy Price Processes Used for Derivatives Pricing & Risk Management', Commodities Now. September. Available online [HREF6]
Curbera, F., Nagy, W.A. & Weerawarana, S. (2001). 'Web services: Why and how', OOPSLA 2001 Workshop on Object-Oriented Web Services, Tampa, Florida, USA, October 2001
Fayad, M.E., Schmidt, D.C. & Johnson, R.E. (1999). Building application frameworks: object-oriented foundations of framework design, 1st edn, John Wiley & Sons, New York.
Schwabe D., Esmeraldo L., Rossi G., Lyardet F. (2001). 'Engineering Web Applications for Reuse' IEEE MultiMedia vol. 8, no. 1, pp. 20-31
Turner M., Budgen D., Brereton P. (2003). 'Turning Software into a Service', IEEE Computer vol. 36, no. 10, pp. 38-44. Available online [HREF7]
Booth D., Haas H., McCabe F., Newcomer E., Champion M., Ferris C., Orchard D. (2004). 'Web Services Architecture' W3C Working Draft, viewed 10 February 2008. Available online [HREF8]