So many words, phrases, and processes that were used in 70s regarding computers have faded to the point that it is likely that few under 50 years of age would have heard of them, yet alone understand their usage. I am referring to terms such as core memory, thrashing, I/O bound, DP, TOS, DOS, boot strap, JCL, punched cards, re-IPLing, LCS, and core dump. This was the era of mainframe computers with orders-of-magnitude less processing power and storage than available today at orders-of-magnitude higher prices. Indeed, as the phrase DP suggested, this was a period of batch processing of data, e.g., payroll or inventory update, versus that of dynamic generation of information.
Back then, one process in particular was extremely important when making decisions about the investment in computer systems. That is, customers would often require that a computer supplier to benchmark it’s various levels of systems, or against competitor systems, to compare the efficiency and adequacy relative to the cost of those systems. In truth, having been actively involved in such activities as an IBM DP Marketing Representative, benchmarking was more often than not a shell game played by the vendors’ System Engineers that tweaked each part of their computer parameters (constraints such as I/O speed, partition size, disk access speed, etc.) to maximize the throughput of a particular a system in favor of each customer’s individual expectations. Getting the customer’s order was often the result of the vendor’s Marketing Representatives and System Engineers working together to set up the customer expectations and then to demonstrate the ability to meet them, respectively, euphemistically referred to as having account control.
With the move from back-office / mainframe computing to that of distributed client/server, the art and science of benchmarking has become that of legends for those still able to remember the good ole days. With seemingly unlimited computer power and communication links, there is rarely an issue today of whether or not the IT architecture will handle the requirements and at what cost. Over the last several decades the investment decision has shifted from costs/power to that of developing / obtaining the software relative to business value. Well, that’s almost totally true. For industries that rely on substantial mobile and remote resources, advancing IT can also be a significant infrastructure and hardware cost as well and therefore worthy of benchmarking. Unfortunately, that continues to NOT be the case for railroads.
As I have noted in other postings on this blog, and in my quarterly publication, Full Spectrum, most of the major railroads in North America have failed to develop a strategic technology plan in sync with a strategic operations plan (Strategic Railroading™). What is not understood, and therefore not appreciated or evaluated by railroads, is the paradigm shift that can be made for these predominantly unscheduled railroads by increasing the accuracy and timeliness of the status of their key assets, including track time, locomotive diagnostics, fuel, crews, and yard occupancy. And, the primary technology to do that is wireless data. So, if companies found it appropriate to benchmark computer systems for the paradigm shift that they made in the 70s by replacing clerks with MIPS, then why are the railroads not doing the same in pursuing their deployment of wireless? There are two points to consider to address this question, i.e., MOTIVATION and PROCESS.
In the case of railroads with 1000s of locomotives and the possibility of incorporating them as mobile nodes on the IT architecture, as a manufacturer would consider fixed nodes, then there is definitely something missing. What is missing is the understanding by railroad management, and suppliers failing to taking a proactive position, of what can be done with IN-TIME data. I am not referring to REAL-time data. The difference between IN-TIME and REAL time is critical in understanding the constraints of using wireless data, versus the seemingly infinite capability of wired links as in a manufacturing environment. To be explained in a future posting, IN-TIME data for train speed and position information in unscheduled operations is no more frequent than every 5 minutes for other than moving block operations. Hence, the railroad technicians that are charged with designing wireless networks can’t help themselves, nor are they held responsible, in making technical decisions which are not related to true business evaluation. Stated simply, technicians will always over design to make sure that they don’t come up short.
As nearly everyone now appreciates with the proliferation of cell phones and laptop computing, wireless is clearly limited in its throughput speed and coverage. It has been an eye-opening experience for those folks that expected that their internet connectively on their cell phone and notebook would match their in-the-office-cubicle desktop performance. There are two primary ways to determine what needs to be done.
1. evaluate every possible wireless-based application as to data requirements and calculate the ultimate throughput requirements. At least one railroad tried this approach several years ago, and the process bogged down in detail thereby insuring nothing would be resolved.
2. evaluate on an 80/20 basis as to evaluating throughput requirements relative to a variety of wireless options recognizing the two key parameters of wireless data parameters. i.e. throughput and coverage. This approach was used a decade ago when I structured such a study that was participated in by the big 4 railroads in the U.S. with oversight by the AAR. The results of that study were used at that point by the AAR to justify the industry’s usage of the 160 VHF spectrum to the FAA. However, that was all the farther it went. Basic details follow.
Developing a Wireless Strategy for a railroad, or for an industry, needs to be pragmatic and adjustable to each railroad’s technical agenda, assuming there is one.
The process is rather simplistic in structure, but a true commitment is required by a railroad’s upper management to provide the players involved with the proper motivation to address the bottom line at the same time. To be brief, the process requires developing a matrix that plays off THROUGHPUT requirements against COVERAGE. For railroads, the THROUGHPUT requirements may include simple categories such as Voice, Monitoring (locomotive diagnostics, shipment status), Out-bound Transactions (PTC targets), Process Control (moving block), and Interactive (M of W activities, in-train management). As to COVERAGE, the categories can be as simple as Terminal, Metropolitan (major cities with multiple railroads), Main Line, and Group (M of W gangs, Trains/Cars). Within each Throughput / Coverage block of the matrix, the possible applications are identified with a pragmatic evaluation of data requirements. This provides the Demand perspective.
The next step is to evaluate the various wireless data options, both commercial and private, as to their ability to service the demand. This is the Supply perspective that results in Wireless Corridors, if you will, that permits structuring a manageable number of wireless strategies based upon business evaluation as to costs vs. value. Such an analysis, in my belief, would have prevented the phenomenal, unwarranted investment in the 220 MHz spectrum that is being made in the name of PTC, even though the railroads are required to spend $100s millions to rebuild the 160 MHz infrastructure as required by the FCC by 2013.
For those small to medium railroads outside of North America that are being slammed with ETCS, and the requirement for GSM, the analysis goes even deeper. That is, as described in other postings on this blog, the use of dark territory (with or without PTC) and the deployment of cost-effective wireless solutions can provide substantially lower capital investments to run a railroad both safely and efficiently.
Bottom Line, railroads should be benchmarking the use of wireless technologies with a pragmatic understanding of both Demand and Supply. Further details of such a process can be obtained by contacting me to discuss individual situations. This is what I provide as a consultant.