HP XP7 Storage - Manual (Page 40)

4 Planning the data path
A data path must be designed to manage your organization’s throughput to the remote site. The
bandwidth, number of ports, and fibre-channel data path configuration that you use help ensure
that your update data arrives at the remote site in a time consistent with your organization’s RPO.
This chapter provides instructions for calculating bandwidth and designing the data path network.
Data path design workflow
To set up a data path, you must establish the following:
•
The amount of bandwidth to move data generated by your host application under all I/O
conditions. See
“Sizing bandwidth ” (page 40) .
•
Ports that can send and receive data. See “Planning ports for data transfer ” (page 45) .
•
The type of cable and number of switches required. See “Cable length and switch requirements”
(page 47) .
•
The configuration that works best for your sites. See “Supported data path configurations ”
(page 48) .
This chapter discusses these topics in detail.
Sizing bandwidth
You purchase bandwidth according to the amount of data that will be transferred from the primary
to the secondary system within a certain amount of time.
If the data path network cannot keep pace with the flow of data, the data is saved in the journal
until additional bandwidth capacity becomes available. If the journal also cannot keep up, the
integrity of the pairs is lost and a new initial copy must be created.
In general, bandwidth is expensive. Increasing capacity to a journal volume is relatively inexpensive.
But the more data that accumulates in the journal, the further the secondary image is from the
production volumes. Therefore, sizing bandwidth is a trade-off between expense and keeping your
secondary volume as consistent as you need it to be with the primary volume.
Five sizing strategies
The following sizing strategies are provided to help you work out an approach to sizing bandwidth.
Be aware that these are not the only strategies you can use.
•
Size bandwidth to peak workload. This results in the smallest difference between data in the
P-VOL and S-VOL. Identify peak workload on the production disks, then add extra capacity
to accommodate packet loss and protocol overhead. RPO is at or near zero when bandwidth
is sized to peak workload.
•
Size bandwidth to peak workload rolling average. The rolling average is less than peak but
more than average. This guarantees that at some point data will accumulate in the journal,
but most of the time it will not. Your system can afford to journal for the amount of time planned
for and still maintain RPO.
•
Size bandwidth to typical workload. When bandwidth is sized to typical write-workload, and
an extended peak workload is experienced, excess write-data is written to journal. This excess
data is delayed for subsequent FIFO transmission to the remote site when network capacity
becomes available. Differential data is proportional to the amplitude and duration of the
workload surge.
•
If you cannot determine a “typical” workload, sizing should be to the average or mean
workload, plus a small compensation for network overhead. In this scenario, excess data in
40 Planning the data path