SOFTWARE The future of software
Jason Maude, chief technology advocate at Starling Bank, discusses agile innovation and why resilience is key
To understand how Starling has eliminated the
I
trade-off between speed and security, we need to look at how software systems are created and maintained, and how the risks are managed. Many organisations attempt to reduce the risk inherent in changing their software by focusing on lowering the probability of errors occurring, be it in the form of bugs, outages, or other occurrences that negatively impact users. In order to do this every time a software
release is planned, a lengthy set of procedures is invoked. These provide a process in which software changes are examined, tested and checked by different groups of people. This can take days or even weeks to complete. Due to the effort this requires, the frequency of releases must be reduced, usually to a monthly or quarterly basis. This does not eliminate errors entirely, of course. Generally speaking, an error discovered in a release will result in a certain degree of panic, meetings to dissect the error, and more steps added to the release procedures to prevent that particular error reoccurring (which of course lengthens the procedure). At Starling we take a different approach to
reducing software change risk. Instead of prioritising the reduction of the probability of errors, we focus on reducing the impact of an error on the users of the system. The best way to reduce error impact in a software system is to make the software correct the errors itself wherever possible. The software will generally
40 |
“The best way to reduce error impact in a software system is to make the software correct the errors itself wherever possible. Whenever an error occurs, our preference for preventing it in the future is to change the software to elminate or lower the impact of the bug”
respond to problems and correct them far more quickly than a human would. Not all errors will be corrected, but, if it can handle 90% of the most common and easily definable cases, then the software’s resilience will be drastically increased. Therefore, whenever an error occurs, our preference for preventing it in the future is to change the software to eliminate or lower the impact of the bug, rather than attempting to lower the probability of such an error occurring by adding a clause to a release procedure. The resilience of the software allows us to
make changes to the system at speed, while simultaneously relying on its security. We release software changes to our production code dozens of times per day, allowing us to deliver new products, services and features rapidly. At the same time, we maintain a high level of reliability and security. We have had no major incidents in the past 12 months. Overall, we trust our software because it does a huge amount of checking and testing for us. The future of software within financial
services is going to require faster changes, with a greater commitment to reliability and security. No organisation will be able to survive if it has to trade off between these two goals. Only with a software system designed with resilience in mind can you survive the fintech revolution.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46
