
Ever wondered what systems are used for mission critical systems? Those that need
99.999% uptime?
This includes air traffic control systems, hospital systems, bank systems, and industrial automation
systems.
As complex as today's servers are, they leave a lot of room for problems including failures of
fans, power supplies, CPUs, Memory, disks, Ethernet ports, fiber channel ports, and others.
In some cases the failure of any of these can bring a workload down with it.
Using VMware's High-Availability feature in vSphere / ESX that can power-on a virtual machine
on a different physical server if its current physical server fails sounds like a fool-proof
plan, but there are some disadvantages:
- The time for ESX to determine the failure
- The time to transition the workload to another physical server
- The time to power on the workload and boot
- The possibility that corruption occurred in the filesystem - requiring a consistency check
- The possibility that corruption occurred at the database layer - requiring a consistency check
- The possibility that corruption occurred at the application layer - requiring a consistency check or manual intervention
As you know, consistency checks can take a
long time, especially for Exchange or for very large databases,
and many have to be performed offline. A simple hardware failure could bring down systems that are critical
to operations, and even with HA, there could be significant downtime.
The solution?
Hardware fault-tolerance.
With NEC's Fault Tolerant Servers, hardware-layer fault-tolerance is achieved, providing instant
fail-over of a workload (including an entire ESX host) with NO interruption.
Recovery after replacing a failed module in the fault tolerant server is non-intrusive and
does NOT affect the workload on the surviving module, even with high-stress workloads.
As an
NEC Hosting Partner, Genesis is proud to be the first to offer high-availability hosted solutions
using NEC's Fault Tolerant servers for critical systems for smaller banks, airports, and businesses
that need robust infrastructure, but
don't have the budget to build-out a local data center room nor manage the physical or virtual infrastructure
necessary for their environment.