HP Insight Recovery Technical White Paper

6
Table 1
: Example of
Applicat
ion Tiers according to
required
l
evels
of availability
In this
hypothetical
example,
a
customer has classified their applications into four “tiers” of
application
availability, labeled Tier 1 (
Top X% Mission Critical) to Tier 4 (Non
-
Business Cr
itical
). This
example illustrates how RTO
and RPO
requirements may vary across application tiers. In this case,
Tier
1 applications have an RTO of less than
2 hours
,
and an RPO of
less than
5 to 10 minutes
.
The point to be observed in this example i
s that not all applications
have the same
DR requirements,
that is
,
some
may
have more demanding requirements for RTO and RPO than ot
hers. Each enterprise
will have
its own set of criteria for defining application tiers, and its own requirements
that are
documented in
SLAs and SLOs
.
Planned
vs. U
nplanned
Disaster R
ecovery
S
cenarios
There are several
use case
scenarios
where DR solutions are
employed, which include
minimizing or
avoiding downtime associated with either
“planned” and “unplanned”
events
. Fo
r example, a
planned
failover
scenario may be initiated i
n anticipation of
an approaching hurricane
, or
ahead of
a
scheduled
maintenance activity
that would affect the entire site
,
such as
a power or cooling
upgrade
.
In planned scenarios,
there is
typical
ly
sufficient time to perform an orderly shutdown
of
services at the
primary site
before initiating the failover of services at a
recovery site
.
In unplanned failover scenarios,
typically
a catastrophic event has already occurred,
and
the
primary
data c
enter
(or some portion thereof)
has been rendered
inoperable. In
these
cases, an unplanned
failover
involving the restoration of services at a
recovery site
will
be
required
. In a
n unplanned
scenario,
there
may not be an
opportunity to
perform
an orderly
shutdown of services at the
primary
site
before initiating the failover.
Best effort, if
necessary
< 20 days
2 to 5 days
< 2 hours
RTO
Best effort, if
necessary
< 5 days
1 to 2 days
< 2 hours
RTO
< 3 days data loss
< 1 day data loss
< 1 hour data loss
< 5 min data loss
N/A
Manual Process
Synchronous
Synchronous
N/A
N/A
Metro area
Metro area
Disaster Recovery (Metro Area Distance) for Local Outage Events
/ Local Disasters
Failover distance
Data replication
RPO
Disaster Recovery (Long Distance) for Major Wide
-
Area Disasters
N/A
N/A
> 150 miles away
N/A
Manual Process
Asynchronous
Asynchronous
Data replication
99.0
99.5%
99.5
99.9
99.95%
Application Availability SLO
< 5 days data loss
< 2 days data loss
< 2 hours data
loss
< 10 min data loss
RPO
> 150 miles way
Failover distance
Tier 3
(Business Critical
Production)
Tier 4
(Non
-
Business Critical)
Tier 2
(Mission Critical
Production)
Tier 1
(Top X% Mission Critical
Production)
Tier Name
Best effort, if
necessary
< 20 days
2 to 5 days
< 2 hours
RTO
Best effort, if
necessary
< 5 days
1 to 2 days
< 2 hours
RTO
< 3 days data loss
< 1 day data loss
< 1 hour data loss
< 5 min data loss
N/A
Manual Process
Synchronous
Synchronous
N/A
N/A
Metro area
Metro area
Disaster Recovery (Metro Area Distance) for Local Outage Events
/ Local Disasters
Failover distance
Data replication
RPO
Disaster Recovery (Long Distance) for Major Wide
-
Area Disasters
N/A
N/A
> 150 miles away
N/A
Manual Process
Asynchronous
Asynchronous
Data replication
99.0
99.5%
99.5
99.9
99.95%
Application Availability SLO
< 5 days data loss
< 2 days data loss
< 2 hours data
loss
< 10 min data loss
RPO
> 150 miles way
Failover distance
Tier 3
(Business Critical
Production)
Tier 4
(Non
-
Business Critical)
Tier 2
(Mission Critical
Production)
Tier 1
(Top X% Mission Critical
Production)
Tier Name