Virtualisation puts meat on compute availability bones

Manufacturers can now go for software-based high availability on their production, CAD and ERP systems instead of buying expensive clusters and specialist multi-redundant hardware. Software-based fault tolerant environment developer Marathon Technologies last month launched an extension to its suite, providing continuous availability on split-site systems with remote servers using patented virtualisation technology. Marathon director of systems engineering Greg Cullen says the earlier limits of single and dual processors have been swept away, and the software now scales to four, eight and 16-way servers. “Our product will compute through failures, with Marathon taking failed components off the configuration to protect applications and the data,” says Cullen. “You can do it in hardware, but it’s expensive and there’s the ongoing maintenance costs. Also, with our approach, users don’t need specialised applications that support hardware redundancy. They can use off-the-shelf software. “We work with Wonderware, Siemens, Rockwell Automation – on manufacturing plants, utilities, energy plants.” Manufacturing clients include GM, Ford, Bosch, VW and Audi – and in the UK, fresh meat products manufacturer Dungannon Meats. Dungannon Meats operates in one of the fastest moving, high pressure retail supply chain environments. Customer orders arrive via EDI early to mid morning on the day of production, and all orders are manufactured for same day despatch to customer depot. The company sought to implement IT systems that would be immune to the potentially devastating consequences of hardware failure. Following a review of requirements Dungannon’s IT partner, Real Time Systems, introduced the Company to Marathon Technologies’ FTvirtual Server as a solution. Marathon’s software protects Windows applications from downtime in the event of server failure by creating a virtual server, running all applications simultaneously on two standard Intel servers. If one of the servers fails, the virtual server continues to operate on the remaining server and users experience no interruptions to IT service.