Hardware requirements

Edit online

Hardware requirements for your IBM Storage Protect solution are based on system size. Choose equivalent or better components than those items that are listed to ensure optimum performance for your environment.

For more information about planning disk devices, see Planning for disk storage.

For more information about planning tape devices, see Planning for tape storage.

The following table includes minimum hardware requirements for the server and storage. If you are using local partitions (LPARs) or work partitions (WPARs), adjust the network requirements to take account of the partition sizes. The figures in the table are based on a data ingestion rate of 14 TB per hour.

Hardware component	System requirements
Server processor	8 processor cores, 3.42 GHz or faster. For example, use a POWER8 processor-based server. 16 processor cores, 2.0 GHz or faster. For example, use an Intel Xeon processor.
Server memory	64 GB RAM.
Network	The following sizing manages approximately 14 TB of data per hour: 10 Gb Ethernet (a minimum of four ports) 8 Gb Fibre Channel adapter (a minimum of four ports) The number of ports depends on the percentage of daily data ingestion to disk storage pools versus tape storage. Use separate Fibre Channel adapters for tape and disk data.
Storage	Disk Based on the amount of data that you are writing to disk, specify the number of disks that you require. Ensure that the sequential input/output (I/O) throughput of the storage area network (SAN) matches the I/O throughput for the network in the previous row. For example, if you must back up 10 TB of data in a four-hour window, the throughput is approximately 700 MB per second. In this case, the server requires a front-end network (client-to-server path) that supports a minimum throughput of 700 MB per second. The back-end SAN (the server-to-storage device path) also must support a minimum throughput of 700 MB per second. To calculate the required disk speed, use the following formulas: `(Total amount of daily data ingestion - amount of daily data ingestion directly to tape) ÷ (Number of hours for daily client backup operations) = Megabytes of data ingestion to disk per hour` `(Megabytes of data ingestion to disk per hour) ÷ (3600 seconds per hour) = Megabytes of data ingestion per second that must be supported by the disk technology` Tape Select the tape technology that best fits your business requirements. For example, use IBM Linear Tape-Open (LTO) or IBM TS1150 tape drives. Ensure that you have sufficient mount points for client backup operations and for migration. For more information about planning tape storage, see Planning for tape storage. For a list of supported tape devices, see IBM® Support Portal for IBM Storage Protect. Tip: To optimize data movement, use LAN-free data movement.
SAN I/O adapters	Segregate disk and tape I/O. For more information about selecting an adapter, see the documentation for Brocade hardware products and for IBM Storwize® storage solutions. Disk Use at least two adapters. Tape Use at least two adapters.

Estimating space requirements for the Operations Center

Hardware requirements for the Operations Center are included in the preceding table, except for the database and archive log space (inventory) that the Operations Center uses to hold records for managed clients.

If you do not plan to install the Operations Center on the same system as the IBM Storage Protect server, you can estimate system requirements separately. To calculate system requirements for the Operations Center, see the system requirements calculator in technote 1641684.

Managing the Operations Center on the IBM Storage Protect server is a workload that requires extra space for database operations on both the hub server and any spoke servers. The amount of space on the hub server for the archive log is larger if the hub server is monitoring one or more spoke servers. Review the following guidelines to estimate how much space your IBM Storage Protect server requires.

Database space for the Operations Center

The Operations Center uses approximately 4.4 GB of database space for every 1000 clients that are monitored on that server. This calculation applies to both hub servers and spoke servers within a configuration.

For example, consider a hub server with 2000 clients that also manages three spoke servers, each with 1000 clients. This configuration has a total of 5000 clients across the four servers. Each of the spoke servers requires 4.4 GB of database space. If the spoke servers are at IBM Storage Protect Version 8.1.2 or later, the hub server requires 8.8 GB of database space for monitoring only its 2000 clients:

(4.4 GB x 2) = 8.8 GB

Database space for managed data

Managed data is the amount of data that is protected, including the amount of data for all retained versions.

For client types that perform incremental-forever backups, the following formula can be used to estimate the total managed data:
```
Front-end + (front-end × change rate × (retention - 1))
```
For example, if you back up 100 TB of front-end data, use a 30-day retention period, and have a 5% change rate, calculate your total managed data by using the following figures:
```
100 TB + (100 TB × 0.05 × (30-1)) = 245 TB total managed data
```
For client types that perform full backups every day, the following formula can be used to estimate the total managed data:
```
Front-end × retention × (1 + change rate)
```
For example, if you back up 10 TB of front-end data, use a 30-day retention period, and have a 3% change rate, calculate your total managed data by using the following figures:
```
10 TB × 30 × (1 + .03) = 309 TB total managed data 
```
Unstructured data, average object size: 4 MB

Structured data, average object size: 128 MB

Unstructured data, number of objects =

(245 TB × 1024 × 1024) / 4 MB = 64225280

Structured data, number of objects =

(309 TB × 1024 × 1024) / 128 MB = 2531328

Total number of objects: 66756608

Managed data cost (1 KB per object) =

(66756608 KB) / (1024 × 1024) = 63.66 GB

Plan for 20% of additional space so that database systems are not at 100% capacity:
```
Database total physical storage requirements =
(managed data space + Operations Center space) × (1.20)
```
For this example, you would calculate the space by using the following figures:
```
(66.33 GB + 8.4 GB) × 1.20 = 76.41 GB
```

Archive log space

The Operations Center uses approximately 18 GB of archive log space every 24 hours, per server, for every 1000 clients monitored on that server. Additionally, for every 1000 clients that are monitored on spoke servers, additional archive log space is used on the hub server. For spoke servers at V8.1.2 or later, this added amount is 1.2 GB of archive log space on the hub server per 1000 clients monitored every 24 hours.

((18 GB x 2) + (1.2 GB x 3)) = 39.6 GB of archive log space

These estimates are based on the default status collection interval of 5 minutes. If you reduce the collection interval from once every 5 minutes to once every 3 minutes, the space requirements increase. The following examples show the approximate increase in the log space requirements with a collection interval of once every 3 minutes for a configuration in which V8.1.2 or later spoke servers are monitored:

Hub server: In the range 39.6 GB - 66 GB
Each spoke server: In the range 18 GB - 30 GB

Allocate archive log space so that you can support the Operations Center without affecting server operations.