A leading technology manufacturer found their UK based central fileserver
to be business critical when the server itself became unreliable due to
lack of diskspace, hardware failures and transient operating system faults.
In order to regain management and user confidence, they approached Preferred
to assist in sourcing and implementing a replacement solution to provide
A number of key requirements were identified, including:
- A highly available Windows 2000 solution to support their 24x7 operation.
- Increase available diskspace by 200%
- Easy deployment
- Simple administration
- The solution must be designed & implemented by an IT Infrastructure Specialist
with a proven record and demonstrable project management expertise.
Preferred suggested the client should consider the HP Assured Availability
solution based on the Marathon Endurance product, this solution consisted
of hardware and software that work together to integrate an array of four
industry-standard intel servers into a single, logical solution. The four
systems are divided into pairs; each pair includes a Compute Element (CE)
and an I/O Processor (IOP). The Windows operating system and applications
run only on the CE's, the IOPs do all the I/O processing.
Because the Endurance solution separates the operating system and applications
in the CE from the I/O processing in the IOP, it improves overall system
reliability by eliminating the most common cause of operating system conflicts.
It also eliminates the need for a RAID controller, because the solution
provides disk-mirroring that equals RAID 1 functionality. Marathon's Endurance
product keeps your applications running nonstop by detecting, isolating,
and computing through failures with ComputeThru technology. Unlike a cluster,
which requires 10 minutes or more of downtime while it performs a failover,
the Endurance system continues processing with no interruption in service,
no loss of application state, and no operator intervention. If a failure
occurs, the users never know it. With the Marathon No Touch Recovery feature,
the system automatically assimilates new components without operator intervention.
When powered on, it resynchronizes, still totally transparent to users.
- With Marathon you get 99.999% uptime without the cost and limitations of a proprietary solution
or the hassles and failover problems of a cluster.
- Off-the-shelf Microsoft‘ Windows‘ server operating systems
- Shrink-wrapped applications - no modifications necessary
- ComputeThru‘ processing - no failure event will stop a transaction in progress
- Load 'n Goˇ, No Touch Recoveryˇ - any replaced component will automatically rejoin the live set-up
- Continuous data access - data is available even through failures of OS or hardware
- Uninterrupted connectivity - duplicate network connections
- OS fault resilience - patented architecture provides extra level of OS fault resilience for off-the-shelf Windows OS
- No need for scripting or cluster API programming
- Disaster tolerance - optional SplitSite‘ capability allows physical separation of redundant systems