Stratus Technologies Inc. has integrated disaster recovery capabilities in its everRun Enterprise software platform...
for Intel-based x86 servers.
Stratus also released a new version called everRun Express for edge installations that need local storage but may lack dedicated IT resources.
It marks the first major upgrade to the everRun platform since Stratus acquired the availability suite in 2012 from Marathon Technologies Corp. Stratus has its roots in fault-tolerant hardware products, including its ftServer line, but the Marathon deal brought it into software. The vendor said its business is split evenly between hardware and software sales.
Asynchronous WAN replication added for virtualized instances
The everRun Enterprise consists of a loadable Linux module installed on two x86 servers, one of which is designated as the primary target for administration and configuration. The upgraded version adds an optional third everRun node for asynchronous replication over a wide area network (WAN).
A two-node primary site supports up to 24 virtual machines (VMs) through the kernel-based Virtual Machine hypervisor. The disaster recovery (DR) target creates up to six snapshots from primary VMs.
Licensing for everRun Enterprise is $12,000 for two primary servers and $5,000 for the additional DR node.
EverRun Enterprise is designed for bootless fault tolerance and in-flight data protection of Windows and Linux applications. Should one node cease to function, applications and VMs automatically fail over to the remaining node.
Previous versions enabled synchronous replication up to three miles but did not support fault tolerance across a WAN. For DR, administrators had to write scripts and configure files using third-party tools from Stratus software partners.
Use cases include cloud-based DR snapshots
Customers use the Stratus One View management console in everRun to designate the primary site to run in fault-tolerant mode and a second remote location for DR mode. Snapshots capture point-in-time backup copies of the production environment. Stratus claims it provides recovery point objectives of less than 30 minutes and recovery time objectives of less than 15 minutes.
Stratus recently demonstrated its Always-On fault tolerant technology for cloud, which is designed to turn OpenStack into a highly available platform without rewriting applications.
"One use case for DR will be taking snapshots from an OpenStack environment for restoration on a third-party public or private cloud, as nominated by the customer," said John Abbott, a founding analyst at The 451 Group.
Jason Anderson, a Stratus Technologies product manager, said integrating DR directly in everRun makes it easier for customers to deploy in environments outside a company's main data center, such as airports, maritime ports and healthcare settings. Stratus Technologies claims it manages 10,000 such customer installations.
"We're trying to provide the mission-criticality of a mainframe, but make it push-button-easy for people that may not be IT specialists," Anderson said.
The new everRun Express version offers high availability, but not fault tolerance. While everRun Enterprise mirrors an application across two servers, everRun Express runs an application on one server. The Enterprise version replicates and synchronizes data in storage and memory, but everRun Express only replicates the data in storage. The Enterprise version ensures that either server can continue if one fails, while everRun Express requires a restart if the primary server goes down. The Express version requires less CPU and memory and runs faster than the Enterprise edition, but the tradeoff is lack of fault tolerance.
Roadmap includes hardware monitoring
The everRun suite provides application monitoring from the operating system up through the application layer. Anderson said Stratus is lining up partners to add diagnostic hardware monitoring and alerts in 2015 to help administrators pinpoint single points of failure.
Stratus joins a crowded vendor market for high availability (HA) DR products at the virtualization layer. VMware, Citrix and Red Hat have developed native capabilities for their respective hypervisors. Cloud infrastructure software stacks such as Amazon Web Services' availability zones and Citrix/Apache CloudStack have emerged as well as software-defined data center platforms, Abbott said.
"There are independent replication providers such as InMage (acquired by Microsoft), PHD Virtual Technologies/VirtualSharp, Veeam and Zerto, which are arguably more closely integrated [than Stratus's everRun] with the virtualization layer and well-positioned to collaborate with cloud-based partners," Abbott said.
Abbott said competition for Stratus will grow as the OpenStack community develops more mature HA capabilities.
Fault tolerance for SMP comes to everRun
Sullivan Group adopts everRun fault tolerance software