The following table estimates the key data points of availability and durability based on the implementation design, Teradata architecture, the solution scale, and the underlying components characteristics:
Number of 8+1 (ESXi) Cliques | Total Number of Nodes (ESXi), Excluding HSNs | Estimated System Availability (MTBF, Availability in %, hours per year) | System Data Loss (MTTDL, 1 chance in X per year) |
---|---|---|---|
1 | 8 | 53.8e3, 99.99%, 1.1hr | 3.7e9, 1 in 439,000 |
2 | 16 | 27.8e3, 99.98%, 1.9hr | 2.4e9, 1 in 278,000 |
3 | 24 | 18.7e3, 99.97%, 2.8hr | 1.7e9, 1 in 203,000 |
4 | 32 | 14.1e3, 99.96%, 3.6hr | 1.4e9, 1 in 160,000 |
5 | 40 | 11.4e3, 99.95%, 4.5hr | 1.1e9, 1 in 132,000 |
6 | 48 | 9.5e3, 99.94% 5.3hr | 9.6e8, 1 in 113,000 |
7 | 56 | 8.1e3, 99.93%, 6.2hr | 8.4e8, 1 in 98,000 |
8 | 64 | 7.1e3, 99.92% 7.0hr | 7.4e8, 1 in 87,000 |
9 | 72 | 6.3e3, 99.91%, 7.9hr | 6.6e8, 1 in 78,000 |
10 | 80 | 5.7e3, 99.90%, 8.7hr | 6.0e8, 1 in 71,000 |
11 | 88 | 5.2e3, 99.89% 9.6hr | 5.5e8, 1 in 65,000 |
12 | 96 | 4.8e3, 99.88% 10.4hr | 5.1e8, 1 in 60,000 |
13 | 104 | 4.4e3, 99.87%, 11.2hr | 4.7e8, 1 in 55,000 |
The estimates specified in the table are created using the following parameters based on similar systems and may not reflect the actual characteristics of the underlying components:
- Host Node (Analytics Database or Management Node) MTBF - 50k hours
- Array Availability - 99.999% with 4 hour critical support (Dell Provided through ProSupport Plus)
- Switch MTBF - 500k hours
- Copper Cable MTBF - 10M hours
- Optical Cable MTBF - 1M hours
- Fallback – On
- Fallback recovery estimate of approximately 8 hours