Riverstone Networks RS 2100 Especificaciones descargar pdf (Pagina 18)

Fibre Channel Over IP

18 Continuous Access and Data Replication Manager SAN Extensions Reference Guide

3. While collecting the data in step 2, determine the values for recovery point

objective (RPO) and recovery time objective (RTO). RPO is a measure of

how much data can be lost due to a problem at the source site and therefore

defines how real-time the solution needs to be. RTO tells you how long to get

the recovery site going and most include data and application failover and

restart. By definition, an RPO of zero meaning no data can be lost, and

requires synchronous replication regardless of the product chosen to perform

the replication. For asynchronous CA EVA, and asynchronous DRM the

design space is an RPO of near zero. Asynchronous CA XP supports an RPO

from near zero to many hours. All three products support an RTO measured

in several minutes to an hour or two.

Once the data has been collected and depending on the RPO:

4. If the RPO is at or near zero, then use the peak write rate and throughput to get

an initial estimate of how much bandwidth is really needed. For some

real-time applications like Microsoft Exchange, increasing the bandwidth

between 2 to 10 times this initial estimate will reduce the impact on all writes

because of waiting for access to the link.

5. If the RPO is much greater than zero, then average the change rate over the

RPO interval, and use this as an initial estimate of the intersite bandwidth.

Note this bandwidth may need to be adjusted up or down depending on a

particular environment and the amount of time needed to flush the last write

of today before starting on tomorrows work.

Note: These calculations have been explained without considering the impact of

compression because its hard to predict how compressible the data is before hand. If it

is determined that all data is compressible at a constant rate, then that ratio can be

used to reduce the effective throughput required from the link.

Tips and techniques

In a large SAN with multiple copy sets, merges, full copies, or normalization can

take an extremely long time. The following techniques can decrease the time

involved in a DRM environment with HSG80-based storage or in a Continuous

Access EVA environment with the EVA family of storage arrays:

■ Set up and normalize all copy sets at the same location with direct Fibre

Channel connections, and then move the remote hardware to the remote site

for normal operations. Note that new copy sets will then normalize at the

slower link speeds.

1 2 ... 13 14 15 16 17 18 19 20 21 22 23 ... 119 120

Sin comentarios

Riverstone Networks RS 2100 Especificaciones Pagina 18