HP Insight Recovery Technical White Paper

5
definition of reasonable
time
varies
according to the
needs of the enterprise and is
typically
specified
in a
Service L
evel Agreement (SLA) between an
IT organization that maintains the applications
and
the end
-
users who
depend on those applications.
An SLA contains a set of objectiv
es
or
re
quir
ements
regarding service
availability. These are
generally
referred to as Service Level Objectives (SLOs).
SLOs may include requirements for uptime (measured in
“number of nines”,
for example
99.95%)
and
may also include requirements for
Recovery Tim
e Objectives (RTO) and Recovery Point Objectives
(RPO).
Recovery Time Objective (RTO) and Recovery Point Objective (RPO)
When
describing
the DR requirements of a given application
workload
, two key
SLOs
are “Recovery
Time Objective” (RTO) and “Recovery
Point Objective” (RPO).

Recovery Time Objective
(RTO) refers
to the required time to recover an application
(“service”)
,
typically
measured as the time that
all
service
s
are restored from an end
-
user perspective.
These
requirements can range from unde
r two hours for mission critical applications
,
to within 1
-
2 days
for less critical applica
tions
,
to within 1
-
2 weeks
for low priority
services
.

Recovery Point Objective
(RPO) refers to the required point
-
in
-
time to which
a
DR solution can
restore
the st
ate of data
for a particular service
,
for example
, how much data loss is acceptable
following a disaster.
These
requirements can range from “close to zero” data loss
(
for high
-
value
financial tr
ansactions
, for example
) to some
small
number of minutes of t
he most recent data
,
to
loss of the past
1
-
2 days
of
data, depending on the criticality
of the data
to
the enterprise.
The ability of
a par
ticular DR solution to meet SLA
objectives will vary according
to several factors,
and their associated costs,
incl
uding
:

P
roperties of
t
he DR solution

S
torage r
eplication techniques employed

Inter
-
site communication link bandwidth and quality
For example,
a
n RPO requirement of “close to zero” data loss
would tend to favor
a
synchronous
data
replication
technique rath
er than
asynchronous replication
–
however
there are advantages and
disadvantages to both modes of replication
.
In a later section of this document (
Insight Recovery and RTO / RPO C
onsiderations
), the RTO of RPO
characteristics of Insight Recovery are d
iscussed.
Levels of Application Availability
When
developing an SLA agreement, a customer may classify their applications
based on
Application T
iers
according to required
levels of availability, as illustrated in Table 1.