Multiple Redundant Application Servers
Live backup computers that take over when
another computer fails
The main computers that host your website can fail at any time. This
is a fact of life. Multiple identical copies of the backup computer run
together. When one fails, several others take over flawlessly. Your visitors
never see a computer fail.
Redundant Load Balancers
In the rare event that a load balancer itself may
fail, there is a secondary load balancer on
Special computer programs are setup to
automatically check the health of servers 60 times
an hour on each service. If a failed server is
detected, an automatic repair routine is run to
bring the server back online. In the event that
the server cannot be brought back, a technician is
automatically notified. Note that in the even of a
failure other servers have taken over the duties
of the failed server and users will not notice a
difference in service.
In the event that a mail server fails, it will
automatically be restarted. In the meantime, a
secondary mail server will automatically take over
duties of the primary mail server. The result is
that users will never see a failed mail server.
The database server where most of the information
for the website is stored has a backup. In the
event that the primary database fails, the backup
database can take over the duties of the failed
Every 4 hour
Backups are made to other servers on-site every
three hours to protect important information.
5 days per
week off-site backup
Backups are taken off-site 5 days per week to
protect data in a catastrophe.
backup hard drives within a server)
RAID stands for “Redundant Array of Independent
Drives” and means that important information is
stored on at least two hard drives simultaneously.
In the event of a drive failure, the surviving
drive automatically takes over the duties of the
failed drive. The server operates as if the drive
never failed. A new drive is then swapped in and
the data is automatically rebuilt.
class hard drives
All hard drives carrying important data are in
RAID format and are also stored on enterprise
class drives. An enterprise class drive costs two
to four times that of a desktop class drive (i.e.
the ones typically used in offices and homes) but
have much higher MTBF. MTBF stands for mean time
between failures and is a manufacturer rating that
estimates the average time between hard drive
In the event that the firewall fails, a backup
firewall is ready to take over the duties of the
In the event that a file server fails, a backup
file server is ready to take over the duties of
the failed file server.
Backup air conditioning systems take over in the
event of a cooling failure.
|Chilled server room dedicated to housing servers
All servers are protected with backup power units.
|For extended power outages, an on-site backup
power generator is available.