Dan's Take

Stratus Announces a New Focus on Edge Computing

The company faces challenges with its new direction.

• By Dan Kusnetzky
• 12/05/2017

Stratus Technologies is moving aggressively to the edge -- to "edge computing," that is. Its flagship ftServer has been updated with the company describes as "a converged operational technology (OT) edge system" combined with remote management services in a "rugged and easy-to-deploy form factor." Stratus believes that this configuration will be ideal for users in industrial automation (IA) or in distributed enterprises deploying Industrial Internet of Things (IIoT) applications.

Stratus has evolved its messages many times in the past, but its technology has largely focused on providing systems and computing environments that are almost never seen to fail; that quietly go to work and don't require extensive expertise in processing virtualization, such as clustering, virtual machine (VM) software, operating system virtualization and partitioning software or tools to manage workload migration. They're self-monitoring, self-healing and extremely reliable.

What does this really mean? The company's ftServer product line is made up of systems with built-in redundancy that are designed to provide 6-nines of availability (99.9999 percent uptime) out of the box. They don't require special high-availability (HA) software, custom applications designed for HA or any of the complexity found in clusters or grids built using standard, off-the-shelf industry standard systems.

Customers can expect to see downtime on the order of 31.5 seconds per year by simply loading their applications on these machines and letting them run. These systems are built on an x86 foundation and execute industry-standard operating systems. If a component fails, another is immediately (that is, in microseconds) pressed into the service. Operations staff are presented a message telling them this happened so that a repair call can be scheduled. If I'm remembering correctly, these systems can be set up to phone home to request services automatically.

Stratus' everRun software provides a software-based computing environment that provides the order of 3.5-nines (99.95 percent uptime) out of the box. Customers deploying everRun can expect to see something on the order of four hours of downtime per year. As with the ftServers, enterprises can simply load everRun and execute their workloads in that computing environment. everRun, by the way, was a product of Marathon Technologies, a company Stratus acquired back in 2012.

Its V-Series fault tolerant systems use a proprietary hardware and system software environment that offers even higher levels of availability (I've seen the company and its customers mention both 7-nines and 8-nines of uptime in their publications and presentations over the years.)

That means customers could expect to see downtime on the order of a few hundred milliseconds to as much as 3.15 seconds per year. This would be the ideal platform for workloads that can never be seen to fail, that are so critical that using a proprietary platform makes sense. Examples might be electronic funds transfer systems, systems to monitor and control power plants, systems to operate emergency services and the like.

The company's challenge over the years has been to overcome the perception that their systems are expensive when compared to industry-standard servers. While it is true that totally redundant systems are going to be more expensive when only the initial purchase price is considered, the overall cost of ownership would be much lower because customers don't have to use complex or expensive clustering software solutions.

Unfortunately for Stratus, purchasing agents often are tasked only to consider the initial purchase price; not the overall cost to the company, the staffing requirements for a complex computing environment or the potential loss when a complex system either fails or slows down to the point that customers become unhappy.

Dan's Take: Shifting Priorities
Stratus has focused its attention on products for extremely critical transaction processing seen in financial services, healthcare, government and emergency services. Over time, these bastions of Stratus users have been increasingly interested in using clustered solutions with industry-standard systems, processing virtualization, workload migration software and other forms of automation technology. They've been told, by other vendors, that the proper use of processing virtualization and automation can provide lower-cost solutions. It isn't clear, however, whether these customers actually get the same levels of service at a lower cost.

Now the company has found another area in which their continuous processing products can be applied: providing HA computing environments at the edge of the network that are easy to install, easy to use and both heal and manage themselves.

What's not yet clear is whether enterprises will become aware of what Stratus is offering, consider what these products could actually do to provide a "never-fail" computing environment for their IoT workloads or industrial computing environments, find out more about these products, and then select them over competitive offerings built upon virtual processing, virtual storage, virtual networks; i.e., a software-defined computing environment.