X
Business

Stratus merges everRun MX and Avance HA to create continous processing powerhouse

Some workloads simply can't be seen to fail or slow down because the results would be disastrous. Stratus, one of the best known suppliers of continuous processing technology, is merging and refining two of its products to extend performance, manageability and scalability.
Written by Dan Kusnetzky, Contributor

Dave LeClair, Senior Director of Strategy and Product Management at Stratus Technologies stopped by to bring me up to date on what the company has been doing with its software-based continuous processing products, everRun MX and Avance HA.

LeClair said that Stratus was going to announce that it is combining its everRun and Avanace HA technology to create a merged product. This new product would be based upon the KVM hypervisor rather than the Xen technology it used in the past. KVM was selected so that the new merged product could offer improvements in performance, manageability and scale.

Snapshot Analysis

To best understand the impact of this announcement, we have to examine what is continuous processing, what does everRun MX do, what does Avance HA do and then we can consider the benefit of a merged product.

 

What is continuous processing?

Continuous processing is a technique that deploys special purpose systems or software that use redundant components to execute workloads continuously even though some components are taken off line for maintenance or have failed. Work is just reassigned to the available components and people using these applications never know that something is amiss.

Stratus is one of the few companies that offer two approaches, hardware and software, that all offer approximately 99.999% uptime (that's less than 30 seconds per month of unplanned down time.)

What does everRun do?

everRun is a software-based continuous processing product that was designed to keep several independent servers, including those using multi-core or multiple processors, working in lock-step to increase reliability and resiliency while preventing application outages. Stratus calls this "ComputeThru technology.

In the case of a system failure, processing simply continues on the other system. I/O is directed away from the failing devices so that end users are not aware that a failure as occurred.

everRun does that by constantly monitoring all components of an application environment and automatically synchronizing data between them. If it appears that a component is beginning to fail, the system administrators are notified and processing continues on the redundant components.

 

What does Avance HA do?

Like everRun, Avance HA also offers continuous processing capabilities.  It supports this type of processing in a different way, however.

Avance harnesses together two systems using a modified Xen hypervisor and powerful monitoring and automation software. Applications are loaded into a virtual machine that executes on both physical host systems. Data and applications are synchronized between the two host servers continuously. If one fails, the other simply keeps on processing. The end users don't see the failure.

What does this announcement mean?

everRun became a Stratus product when the company acquired Marathon Technologies in September 2012. Critics pointed out that the company then had two products that used different approaches and solved the same problem. They pointed out that it would be very difficult for Stratus to explain which product customers should use and why.

While this acquisition created a little bit of a conundrum for Stratus. It now had two software-based continuous processing solutions, each having unique features, but both offering very powerful tools to support continuous processing environments.

The challenge was how those features could be combined in the same product. The next issue was how to create a migration path for customers using those two different approaches that would lead to a new unified product without causing problems for the customers.

It is clear that after looking at the problem since September 2012, Stratus developed a new product having the best features of each of the older product. Furthermore, this was a time that Stratus could move to a newer, more manageable, more scalable virtual machine technology.

While the company tells me that this release is designed to deliver "mainframe levels of availability" to its enterprise customers, I have no problem visualizing Stratus offering the same capability to service providers allowing them to offer "continuous processing clouds" at some time in the future.

If your company has applications that simply can't be seen to fail, Stratus has several options that would be well worth the effort to examine.

Editorial standards