One of the hallmarks of successful information system architectures is their longevity. Luc Hohmann’s insightful 2005 book Beyond Software Architecture carries this notion in its subtitle: Creating and Sustaining Winning Solutions. Coming back to school from their internships in industry, our students often comment on how old some of the information systems they worked with that summer were. In times where the media keep telling us that current technology will be obsolete in six months, these old systems must indeed appear anachronistic; fossils from the times when people wrote their programs in COBOL on green screens.
Older, so called ‘legacy’ systems or applications survive for any number of reasons. Some of these reasons are not very desirable; for instance, the lack of agility of an organization, users’ unwillingness to learn new things, or the conservative power of those who have sold their hearts to the old system. Yet good system architectures are good precisely because they have been designed to accommodate the vagaries of a changing world; i.e., they can be adapted and extended so that they do indeed ‘live long.’ So long, in fact, that they may survive their original designers.
Still, regardless of how good systems serve their users, eventually, when the cost for fixes, adaptions and extensions becomes too high or when new technologies offer opportunities for providing entirely new services or significant efficiencies and speed gains, it is time to either buy, rent or build something new.
This is the story of one such rebuilds. The system in question is www.teachengineering.org —TE— a digital library of K-12 engineering curriculum that was built from the ground up with established technology and which for 13 years enjoyed lasting support from its growing user community and its sponsors. These 13 years, however, cover the period during which smartphones and tablets became commonplace, during which the Internet of Things started replacing the Semantic Web, during which NoSQL databases made their way out of the research labs and into everyday development shops, during which we collectively started moving IT functions and services into ‘the cloud,’ and during which computing performance doubled a few times, yet again. Alongside these technical developments we saw, certainly since the last five years or so, a rapidly growing emphasis on usability and graphic design, partly because of the need to move applications into the mobile domain, partly because of the need and desire to improve both ergonomics and aesthetics. During this same period, TeachEngineering’s user base grew from a few hundred to more than 3 million users annually, its collection size quadrupled, it went through several user interface renewals, and significant functionality was added while having an exemplary service record, and it enjoyed continued financial support from its sponsors. All of this took place without any significant architecture changes. In Hohmann’s terms, it was indeed a ‘winning architecture.’
Yet, although the system architecture could probably have survived a while longer, it started to become clear that with the newer technologies, better and newer services could be developed faster and at lower cost, that moving most of its functionality into the cloud would both boost performance and lower maintenance cost, and that the system’s resource and code footprint could be significantly reduced by rebuilding it on a different architecture, with different and more modern technology. And, of course, with a new architecture some of the mistakes and unfortunate choices built into the old architecture could be avoided.
In this monograph we provide a side-by-side of this rebuild. We lay out the choices made in the old architecture —we refer to it as TE 1.0— and compare and contrast them with the choices made for TE 2.0. We explain why both the 1.0 and 2.0 choices were made and discuss the advantages and disadvantages associated with them. The various technologies we (briefly) describe and explain can all be found in traditional Information Systems Design & Analysis textbooks. However, in the traditional texts these technologies are typically presented as a laundry list of options, each with advantages, disadvantages and examples. Rarely, however, are they discussed in the context of a single, integrated case study and even rarer in an evolutionary and side-by-side —1.0 vs. 2.0— fashion.
The two systems, TE 1.0 and TE 2.0, both share as well as use alternate and contrasting technologies. Both groups of technologies are discussed, demonstrated and compared and contrasted with each other and the reasons for using them and/or replacing them are discussed and explained.
Along with the discussion of some of these technologies, we provide a series of small cases from our TeachEngineering experience, stories of the sort of things system maintainers are confronted with while their systems are live and being used. These range from strange user requests to Denial of Service attacks and from having to filter out robot activity to covert attempts by advertisers to infiltrate the system. For each of the technologies and for most of these case histories we provide —mostly on-line— exploratory exercises.
This text is meant as a case study and companion text to many Systems Analysis & Design textbooks used in undergraduate Management Information Systems (MIS), Business Information Systems (BIS) and Computer Information Systems (CIS) programs. The US counts about 1,300 (undergraduate + graduate) such programs (Mandiwalla et al., 2016). These texts typically contain short descriptions of technologies which give students some sense of what these technologies are used for, but do not provide much context or reflection on why these technologies might or might not be applied and what such applications actually amount to in real life. As a consequence, students, having worked their way through these textbooks and associated courses will have had little exposure to the reasoning which must take place when making choices between these technologies and to what goes into combining them into working and successful system architectures. It is our hope that this Tale of Two Systems (pun very much intended) will help mitigate this problem a little.
Although instructors of Systems Analysis and Design courses typically have themselves experience in architecting and building systems, setting up demonstration and experiential learning sites equipped with cases where students can explore and practice the technologies discussed in textbooks and coursework can be daunting. Fortunately, technology-specific interactive web sites where such exploration can take place are rapidly becoming available. Examples of these are w3schools.com, sqlfiddle.com, dotnetfiddle.net, codechef.com and others. In the exercise sections of our text (all marked with the symbol) we have done our best to rely on those publicly available facilities where possible, thereby minimizing set-up time and cost for instructors as well as students.
In cases where we could not (yet) rely on publicly available services, we have included instructions on how to set up local environments for practicing; most often on the reader’s own, personal machine.
This book does not (yet) contain lists of test and quiz questions or practice assignments for two reasons. First, we believe that good instructors can and are interested in formulating their own. Second, we have not have had the time to collect and publish those items. However, we very much do invite readers of this text to submit such items for addition to this text. If you do so (just contact one of us), we will take a look at what you have, and if we like it we will add it to our text with full credits to you.
- Hohmann, L. (2005) Beyond Software Architecture. Creating and Sustaining Winning Solutions. Addison Wesley.
- Mandiwalla, M., Harold, C., Yastremsky, D. (2016) Information Systems Job Index 2015. Assoc. of Information Systems and Temple University. http://isjobindex.com/is-programs