Before you deploy Risk Analysis for Sensitive Data Governance, use these guidelines to select the appropriate hardware for your index server and content analyzer computer.
Hardware Specifications for the Index Server
Use these guidelines to select the appropriate hardware for your index server.
Considerations
Size up to the larger specification if either of the following cases applies to you:
Your environment is between two sizes
You are using one Index Server for both Exchange data sources and Exchange backups
If you are using one Index Server for both files and email messages, use the size that matches the larger source. For example, if you have 40 TB of file source data and 15 TB of email source data, use the specification for a medium-sized server.
The following options affect the performance of the Index Server:
Custom entities that include multiple regular expressions
Data classification plans that include the optical character recognition (OCR) option or the pre-generated previews option
File Data and Email Messages
Component | Medium | Small |
|---|---|---|
File source data size per node | 160 TB | 80 TB |
File objects per node (an estimate based on an average file size of 2 MB and on the assumption that 50 percent of documents are eligible for content indexing and there is more than one version of the file) | 40 million | 20 million |
Email source application size | 15 TB | 5 TB |
Email objects per node (an estimate based on an average message size of 100 KB. Messages with attachments are considered a single object) | 150 million | 50 million |
CPU | 16 cores | 8 cores |
RAM | 32 GB | 16 GB |
Index disk space (SSD class disk recommended) | 6 TB | 3 TB |
Note: If the size of backed up data exceeds the prescribed limits, then you must configure dedicated index servers for email and file data based on the specifications as mentioned in the Specifications for Dedicated Servers for File Data and Specifications for Dedicated Servers for Email Messages sections.
Specifications for Dedicated Servers for File Data
Component | Large | Medium | Small |
|---|---|---|---|
Source data size per node | 320 TB | 160 TB | 80 TB |
Objects per node (estimated based on an average file size of 2 MB and on the assumption that 50 percent of documents are eligible for content indexing and there is more than one version of the file) | 80 million | 40 million | 20 million |
Objects per node (estimated for live scan based on an average file size of 2MB) | 160 million | 80 million | 40 million |
CPU or vCPU | 16 cores | 16 cores | 8 cores |
RAM | 64 GB | 32 GB | 16 GB |
Index disk space (SSD class disk recommended) | 12 TB | 6 TB | 3 TB |
Specifications for Dedicated Servers for Email Messages
Component | Large | Medium | Small |
|---|---|---|---|
Source application size | 25 TB | 15 TB | 5 TB |
Objects per node (an estimate based on an average message size of 100 KB. Messages with attachments are considered a single object) | 250 million | 150 million | 50 million |
Number of mailboxes (based on an average mailbox size of 5 GB, and an average of 50,000 messages per mailbox) | 5000 | 2000 | 400 |
CPU | 16 cores | 16 cores | 8 cores |
RAM | 64 GB | 32 GB | 16 GB |
Index disk space (SSD class disk recommended) | 10 TB | 6 TB | 2 TB |
Hardware Specifications for the Content Analyzer
Content analyzer computers detect personally identifiable information (PII) in the data. Some Risk Analysis environments require dedicated content analyzer computers.
Review the following use cases to determine if you need dedicated content analyzer computers:
You need to identify machine learning-based entities. For example, Address, Contextual Date, Location, Money, Organization, Person, and Time are entities that are identified by using machine learning. For a complete list of built-in entities, see Personally Identifiable Information.
You want to configure parallel processing. You can add multiple content analyzer computers, and then create a separate data classification plan for each content analyzer.
Before you install the Content Analyzer package, use the following guidelines to select the appropriate hardware for your content analyzer computers.
Use case | CPU | RAM | Disk space |
|---|---|---|---|
Machine learning-based entities | 32 cores | 64 GB | 2 TB |
Parallel processing | 16 cores | 32 GB | 1 TB |
Tip: To distribute the processing load, you can install the Content Analyzer package on multiple computers.