Fault tolerant servers from Stratus have redundant components and can ensure annual downtime is less than 6 minutes
Stratus Technologies opened its regional office in Dubai, UAE in the month of July 2011. Exclusively focussed on the high end mission critical enterprise infrastructure for the last thirty years, the vendor is known for its fault tolerant server systems. Over the last few years the vendor has seen its regional sales surging ahead and in order to support its partners better it decided to set up an office in Dubai.
For Stratus the partner is the first point of contact for any customer in the region. The selection of a partner is based on whether they are able to deliver the same level of support assurance to the end-customer that Stratus gives to its partners. Says Maged Shaker, Partner Director, Stratus Technologies Middle East, “We select our partners based on how they will be able to support their customers achieving the same level of responsibility we put on ourselves for our customers.”
Stratus’ products are all about supporting continuous availability of applications for its customers. For mission critical operations that cannot afford to be non-responsive or unavailable for any period of time, where business loss is estimated in millions of Dollars for even a short period of minutes, the Stratus 2600, 4500, 6310 server portfolio is usually the appropriate solution.
In more technical language, the Stratus server portfolio fits in where a customer requires a continuously available platform with uptime of at least 99.999%. Across an annual calendar this implies that the server platform cannot have an aggregated downtime of more than 5.3 minutes across the year and should be available for rest of the time without any deterioration of performance.
For Stratus, a partner is a business entity that has sold its products because of an application or solution they are supporting. “If a customer wants five nines they will approach us. If they do the sale through us they become a partner,” says Shaker. The regional office is meant to provide service and training to such new partner entrants.
“The whole idea of opening up an office is to provide more direct support for the partner and working with them to ensure a solution is up and running properly.”
Stratus has a three-fold approach for achieving continuously available uptime in its fault tolerant range. First, every component inside the system is replicated; second, both pairs of components are working simultaneously at the same time and are synchronised; and third, more than 500 internal operations are monitored at the same time through systems software.
“We realise that hardware fails and the only way to tolerate failure is to replicate everything. We also don’t react to a problem. We proactively fix the problem before it happens”.
Unlike high availability server clusters which are designed to recover from failure, continuously available systems are engineered to prevent server failures from happening in the first place. A continuously available server platform from Stratus is also called a fault tolerant one. The primary hardware components are built around Intel based technology and is of standard design. Options for the operating system are also industry standard platforms including Microsoft, Red Hat and Linux.
Within the structure of a single fault tolerant server is the functional equivalent of two conventional servers, which simultaneously process the same data in the same way. The two halves of a fault tolerant server are so closely coupled that the operating system and application detect only a single processing unit. So even though the specifications of a fault tolerant server appear to be duplicated and hence redundant, for the application and processing bus, it is just one of everything.
The key enabling technology in fault tolerant systems is the processor synchronisation lockstep design, also known as processor determinism. Intel is the only processor vendor to make this available in its chipsets. Stratus provides fault detection and isolation hardware logic to compare the results from the two processor channels. During fully loaded operations, application software and middleware remain unaffected if a duplicate component is removed from the server.
During normal operations hardware and software errors are flagged to the partner and Stratus managed support network. This allows possible points of failure to be detected early in the cycle. If there is component failure, the backup component within the server kicks in. The system performance remains secured and replacements and repair can be completed without affecting end user business performance.
With Stratus’ new office located at the hub of the region, Shaker expects support levels for partners to be enhanced to a new level.
– TEA Computers SA, Egypt
– Bader Al Mulla and Bros, Kuwait
– SOL Group, Kuwait
– Mideast Data Systems, Lebanon
– International Technology Services Limited, Libya
– Total Technical Triangle, Saudi Arabia
– CertSoft, UAE
– Emirates Business Machines, UAE
– International Technology Services Limited, UAE
Stratus Fault Tolerant servers
– Intel Xeon Processor, 4 core and 6 core per processor
– Intel Quick path architecture
– Stratus chipsets for core error detection, fault isolation, synchronisation logic and lockstep architecture
– Memory 4/16 GB, 4/48GB, 8/96GB DDR3
– Internal SAS drive 146/300 GB + 1 TB
– Upto 72 RAID expansion slots
– Operating system Microsoft Windows Server 2008 R2 with Hyper-V virtualization
– Operating system Red Hat Enterprise Linux 5
– Operating system VMWare vSphere 4
– Model type 2600, entry level for remote locations
– Model type 4500, scaleable for multitasking
– Model type 6300, enterprise server for mission critical computing