Conrail Inc, the $3,500m freight railroad company serving the north eastern territory of the US since 1976, is using a parallel system to run its 12,000 route miles and the 3.9m loaded shipments it transports anually efficiently. The company always had large amounts of data, but was not making full use of it, said James Horgan, director of customer logistics in the Transportation Department. It also had an $18m scheduling system used by the customer service division, but Conrail’s logistics department wasn’t taking advantage of the data it contained. As a user the, company says it is more interested in what the technology available can do rather than the fact that it is using a parallel system. It has a central shipment relational database using an AT&T Corp Teradata DBC/1012 which contains cost information and an AT&T System 3600 that together hold 74Gb of data with a maximum storage capacity of 225Gb. It uses GQL graphical query language as the enterprise user access tool. The company pulls operational information from the database which contains about 18 months’ worth of data on all Conrail shipments and schedule activities and is strategic in managing the firm’s policy objectives. Conrail tracks 60,000 individual train cars daily and records 25 to 30 events for each car. The firm requires the database to have capacity to store five years of information and the ability to link key elements of shipment data from a variety of sources and to structure the data to support timely data retrieval. The system has enabled the company to improve customer service by providing more up-to-date information on the status of the customers’ freight at any one time. Tighter schedules have been established and customer benefit from a one-stop billing and query solving system. The system can be used to determine how long cars spend standing in the yard waiting for a customer to load his freight as it is not uncommon for customers to make ‘just in case’ planning their policy. In other words they order more cargo trucks than are needed, but from different companies, just in case one company fails to arrive on time. The Conrail database system should eventually help to avoid the need for customers to do this because if Conrail trucks are delayed for some reason, it hopes to develop a system that enables it to track where other trucks are and provide an alternative. The direct results of using the system are a cut of 8% in transit time and an on-time performance increase of up to 95.5%. Could not have done without it Timely self-evaluation has also been brought in and Conrail has retained and improved its market share. Horgan says, No, this is not all because of the massively parallel computer, but we could not have done without it. In the future the company plans to tie in financial information to the system and measure performance in terms of profit. It also plans to use the data to drive the firm’s strategy implementation and is currently working on a five-year plan to implement this strategy. In the future Conrail anticipates scope for more work with customers and making fuller use of vehicles that have finished a delivery and currently spend time idle and unproductively, waiting for their next load. We are going to have more demands for the information and we are very dependent on the technology to do that, he said. He also added that he would like to see better analytical tools available in the commercial market. The GQL language’s role is to bring information that is needed back from the search, but the raw data retrieved does not answer questions the company has and it is this questioning process on the data that takes up time says Horgan.
