Wednesday, October 1, 2014

Business Complexity: ERPs, SOAs and MSAs

Business Complexity: ERPs, SOAs and MSAs

Cynthia Rettig’s article “The Trouble with Enterprise Software,” while ostensibly about implementing various types of management software and IT projects, is also a much deeper meditation on the nature of complexity as it applies to business.[1] As the amount of data and processes that a business has to manage has risen, so has the complexity of the management process. While evolutionary algorithms, machine learning, and data mining may mitigate this complexity to some degree, Rettig correctly notes that these more advanced tools are used only by a minority of firms, the majority of which use IT systems that may be a decade old or older.

Service-oriented architectures (SOAs) have risen to fill this void through the application of modularity: essentially programs within programs which serve atomic functions and can be swapped out or modified without affecting the rest of the program. This trend has progressed even further with micro-service architectures (MSAs), which are even more finely grained, and may provide just a single function (like converting currency or looking up zipcodes) within a larger program.[2] However, even the components of IT development converge to models of minimum complexity, the underlying software architecture to make all these components work in tandem will stay as complex as ever. To add to the problem, development of these micro-service architectures (and APIs) is conducted by a single person, rather than a team, leading to a proliferation of programming styles and conventions, some of which may not work with others.

However, enterprise-level software development is far from dead. According to a recent article in Forbes magazine, enterprise software IPOs and venture capital financing have both risen over the last year.[3] The article also pays homage to the increased role that cloud computing will play, especially in data analytics. All of these changes are underpinned by software that can run equally well on any device (especially tablets, for which shipments have outstripped notebook and desktop PCs since 2012). The result is an “internet of things”, where almost every object has an online presence, and will be tracked in some database, somewhere. To capitalize on these exponentially growing data, firms will need to invest in ever more robust technological solutions, most likely form third-party providers. The massive ERP type systems of the past, developed for a single company, will be primarily reserved for very large firms which have complex supply chains, but don’t need to respond to rapidly changing consumer demand.

[1] The Trouble with Enterprise Software. Cynthia Rettig. MITSloan Management Review.
[2] Through microservices, a renewed push for simplicity and IT minimalism. Joe McKendrick. <>
[3] Enterprise Software and IPO Lessons Learned from Mary Meeker’s Internet Trends 2014 Presentation. Forbes. <>

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.