The best of Nonstop? The theme of this issue requires reflection. Nonstop has enjoyed many successes over the 50 years. Arguably the first system designed to be fault tolerant. Arguably the first massively parallel architecture allowing near linear scalability. One of the first systems to discover new systems in a network with Expand and having an active routing architecture. One of the first to use a fabric, instead of an I/O bus interconnect. Of course the original company culture was legendary. There are many candidates for ‘best’.
It is a business adage that ‘your greatest strength is your greatest weakness’. Along those lines I am inclined to nominate our Operating System (OS). The original Guardian and now the Nonstop OS, including OSS is a thing of beauty. Designed from the ground up on the belief that everything on the system would, at some point, fail. It was a genius development idea instilling the developers to work through what happens when a particular component is unavailable. We know it as ‘fast fail’ and if anything even appears wrong our failover is so bulletproof, the OS would rather fail over to a known good component rather than trust a flakey one. The integrity of the OS, the applications and the underlying data has always been paramount. Nonstop creates checks for all data movement, ensuring the data remains pure.
It is a unique OS which provides great transaction performance and message switching performance. Batch processing and intense compute applications are not the best fit for a Nonstop system. Near real-time is the sweet spot. Those who understand the architecture and OS understandably fall in love with it. The weakness is that most in the industry do not understand it. New applications and services aren’t developed on Nonstop because it is unique. CIO’s wonder about the ‘black box’ that simply ‘runs’ and why it can’t be replaced with Windows or Linux system? We enjoy a high degree of security since it is a system unknown to hackers so no virus, malware or worms to deal with (knock on wood). The uniqueness of Nonstop has prevented it from being a de facto industry standard. It always ‘appears’ expensive since there are very few actual competitors and to approach the availability, scalability and integrity of a Nonstop you must combine several vendors. Pick an OS – Windows or Linux and add Oracle to it with the high-availability options, add in failover capabilities along with a second system and transaction management and monitoring facilities. All that combined will not run like a Nonstop. You have created a system that specialists must manage and one that requires the application of security fixes likely on a weekly basis. You also need to coordinate between the various hardware/software/database vendors when things go wrong (and they will).
For me then, the best of Nonstop remains the OS and everything associated with it, which is to say Nonstop itself. Thanks to the many software partners and application providers who have embraced this wonderful, unique platform. They get it, as do our many long-standing customers who have trusted Nonstop over these many years. Thank you.
Be the first to comment