Predictive Maintenance for IT Operations requires the medium
package. Details are shown in the figure and tables below. Each worker node must have:
- 768 GB memory
- Thirty-two cores per socket
- More than 6 TB SSD capacity per node
- One dual-port 25 GbE NIC
This information is based on coarsely grained guidance from Dell Data Analytics baseline performance testing.
In the figure below, a 48-port S3148 management switch is used to manage all the nodes.
Component | Estimate |
Daily data volume (DDV) | 500 GB |
Compression factor | 0.15 |
Metadata size factor | 0.35 |
Data retention hot (DRH) | 5 days |
Data retention warm | N days |
Maximum volume per indexer | 300 GB |
RAID level | RAID 0 |
SSD (available) per node | 6 TB |
Calculation | Value |
Volume calculation | ((0.15 + 0.35) * DRH * DDV)/(indexers)) |
Volume per indexer | 160 GB [Assuming eight indexers] |
Storage pool requirements | 1.25 TB |
Storage requirements after 3x redundancy | 3.75 TB |
Available SSD capacity | 24 TB |
For additional sizing tools and information, see: