Here are the 9 disciplined practices we follow:
1) We use open source and otherwise standard-based software at all levels of our server environment. We do this so that we are not overly dependent on any one vendor, or any one platform, or any one coding environment. We can mix and match leading solutions, deploying the right capacity at an appropriate value point.
2) We use standard network and server equipment in our racks as well; again, so that should we need to rapidly replace or expand equipment, the solutions are neither unusual, nor in short supply.
3) We deploy both horizontally and vertically scaled network systems. The technical wizards that run our server racks can explain this in engaging detail––but from the CEO perspective, this means that we can isolate bottlenecks to very small focal points. Because the focal point is small, our adjustment can be precisely (and cost-effectively) targeted. Once cleared, the rest of the system simply hums at a powerful ROI.
4) We pay attention to the secondary systems. Case in point: we have learned more than we ever wanted to learn about both power and cooling. These days, watts per rack and cooling capacities seem to have as much to do with up-time as the primary server solutions.
5) We test our failure recovery procedures. Too many software providers don't actually do this, which is highly dangerous.
6) We track down every possible single point of failure that we can identify. A case in point: some years ago one of our monitoring systems generated an alert that we never received because our paging network (large, national provider) was temporarily down. So, now we have multiple paths for alerts and messages; just as we do for power, cooling, servers, network switches, routers and on and on.
7) Visionlink staff (not a third party), monitor every part of our network, server, and backup systems. I'm not going to say here how many pieces of this puzzle we watch over, but it numbers in the hundreds. The technical team knows the operational parameters of just about everything--all the time.
8) This point is the most important and it's about the people. We have a technical team wise enough to understand that reliability and data integrity is everything to our customers. Our staff doesn't fall prey to the glitz and glamor of the newest and fanciest solution unless it is truly up to par. They do not deploy new technology into our production environment just because it's new. That would be fun––until it fails.
9) Lastly, while number 8 above is completely true, it is also important to understand that we release new software enhancements every 10 days, and that we regularly and fundamentally improve our network and server technologies. Change is necessary to give our customers the best technologies available. We just do change in a disciplined and careful manner for the best possible results.