Skip to content

Commit 6bd6d2f

Browse files
leto-bbqleto-bbq
authored
add resource configuration recommendations in table ainode deployment (#1141)
1 parent c68df2a commit 6bd6d2f

18 files changed

Lines changed: 1000 additions & 409 deletions

File tree

‎src/.vuepress/sidebar/V2.0.x/en-Table.ts‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -58,7 +58,7 @@ export const enSidebar = {
5858
children: [
5959
{ text: 'Deployment form', link: 'Deployment-form_apache' },
6060
{ text: 'Obtain IoTDB', link: 'IoTDB-Package_apache' },
61-
{ text: 'Database Resources', link: 'Database-Resources' },
61+
{ text: 'Database Resources', link: 'Database-Resources_apache' },
6262
{ text: 'Environment Requirements', link: 'Environment-Requirements' },
6363
],
6464
},

‎src/.vuepress/sidebar_timecho/V2.0.x/en-Table.ts‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -58,7 +58,7 @@ export const enSidebar = {
5858
children: [
5959
{ text: 'Deployment form', link: 'Deployment-form_timecho' },
6060
{ text: 'Obtain TimechoDB', link: 'IoTDB-Package_timecho' },
61-
{ text: 'Database Resources', link: 'Database-Resources' },
61+
{ text: 'Database Resources', link: 'Database-Resources_timecho' },
6262
{ text: 'Environment Requirements', link: 'Environment-Requirements' },
6363
],
6464
},

‎src/UserGuide/Master/Table/Deployment-and-Maintenance/AINode_Deployment_Upgrade_timecho.md‎

Lines changed: 24 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -87,7 +87,30 @@ sha512sum timechodb-2.0.6.1-ainode-bin.zip
8787
### 2.3 Environment Requirements
8888

8989
- Recommended operating environment: Linux, macOS.
90-
- TimechoDB Version: >= V2.0.8-beta.
90+
- TimechoDB Version: >= V2.0.8.
91+
92+
#### 2.3.1 Resource Configuration Recommendations
93+
94+
> Note: The resource configuration recommendations in this section apply only to **model inference tasks**. Guidelines for model training tasks will be provided in subsequent releases.
95+
96+
The following are baseline resource configurations for model inference running on a single NVIDIA RTX 4090 (24 GB VRAM). For model inference on AINode, overall throughput can be improved by horizontally scaling the number of GPUs. It is generally recommended to deploy servers with 1, 2, 4 or 8 GPUs.
97+
98+
Specifications of inference tasks used in benchmark tests:
99+
- **Univariate inference**: Historical sequence length: 2880, prediction length: 720
100+
- **Covariate inference**: Historical sequence length: 2880, prediction length: 720, with 20 known covariates
101+
102+
| Number of GPUs (NVIDIA 4090, 24 GB VRAM) | Recommended CPU Cores | Recommended Memory (GB) | Supported QPS for Univariate Inference | Supported QPS for Covariate Inference |
103+
|------------------------------------------|-----------------------|-------------------------|-----------------------------------------|---------------------------------------|
104+
| 1 GPU | 16 cores | 24 GB | 100 | 10 |
105+
| 2 GPUs | 32 cores | 48 GB | 200 | 20 |
106+
| 4 GPUs | 64 cores | 96 GB | 400 | 40 |
107+
| 8 GPUs | 128 cores | 192 GB | 800 | 80 |
108+
109+
**Notes**:
110+
- The CPU and memory configurations above follow this general rule: allocate 16 CPU cores per GPU, and set system memory equal to GPU VRAM at a ratio of 1:1.
111+
- The throughput figures are benchmark references. Actual performance may vary depending on model type, data complexity and deployment environment.
112+
- The throughput of univariate and covariate inference shall be evaluated separately as required, and the two values cannot be summed directly.
113+
91114

92115
## 3. Installation, Deployment, and Usage
93116

‎src/UserGuide/Master/Table/Deployment-and-Maintenance/Database-Resources.md‎

Lines changed: 3 additions & 201 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,6 @@
1+
---
2+
redirectTo: Database-Resources_apache.html
3+
---
14
<!--
25
36
Licensed to the Apache Software Foundation (ASF) under one
@@ -18,204 +21,3 @@
1821
under the License.
1922
2023
-->
21-
# Database Resources
22-
## 1. CPU
23-
<table style="text-align: center;">
24-
<tbody>
25-
<tr>
26-
<th rowspan="2">Number of timeseries (frequency&lt;=1HZ)</th>
27-
<th rowspan="2">CPU</th>
28-
<th colspan="3">Number of nodes</th>
29-
</tr>
30-
<tr>
31-
<th>standalone</th>
32-
<th>Dual-Active</th>
33-
<th>Distributed</th>
34-
</tr>
35-
<tr>
36-
<td>Within 100000</td>
37-
<td>2-4 cores</td>
38-
<td>1</td>
39-
<td>2</td>
40-
<td>3</td>
41-
</tr>
42-
<tr>
43-
<td>Within 300000</td>
44-
<td>4-8 cores</td>
45-
<td>1</td>
46-
<td>2</td>
47-
<td>3</td>
48-
</tr>
49-
<tr>
50-
<td>Within 500000</td>
51-
<td>8-16 cores</td>
52-
<td>1</td>
53-
<td>2</td>
54-
<td>3</td>
55-
</tr>
56-
<tr>
57-
<td>Within 1000000</td>
58-
<td>16-32 cores</td>
59-
<td>1</td>
60-
<td>2</td>
61-
<td>3</td>
62-
</tr>
63-
<tr>
64-
<td>Within 2000000</td>
65-
<td>32-48 cores</td>
66-
<td>1</td>
67-
<td>2</td>
68-
<td>3</td>
69-
</tr>
70-
<tr>
71-
<td>Within 10000000</td>
72-
<td>48core</td>
73-
<td>1</td>
74-
<td>2</td>
75-
<td>Please contact Timecho Business for consultation</td>
76-
</tr>
77-
<tr>
78-
<td>Over 10000000</td>
79-
<td colspan="4">Please contact Timecho Business for consultation</td>
80-
</tr>
81-
</tbody>
82-
</table>
83-
84-
> Supported CPU models: Kunpeng, Phytium, Sunway, Hygon, Zhaoxin, Loongson
85-
86-
## 2. Memory
87-
<table style="text-align: center;">
88-
<tbody>
89-
<tr>
90-
<th rowspan="2">Number of timeseries (frequency&lt;=1HZ)</th>
91-
<th rowspan="2">Memory</th>
92-
<th colspan="3">Number of nodes</th>
93-
</tr>
94-
<tr>
95-
<th>standalone</th>
96-
<th>Dual-Active</th>
97-
<th>Distributed</th>
98-
</tr>
99-
<tr>
100-
<td>Within 100000</td>
101-
<td>2-4G</td>
102-
<td>1</td>
103-
<td>2</td>
104-
<td>3</td>
105-
</tr>
106-
<tr>
107-
<td>Within 300000</td>
108-
<td>6-12G</td>
109-
<td>1</td>
110-
<td>2</td>
111-
<td>3</td>
112-
</tr>
113-
<tr>
114-
<td>Within 500000</td>
115-
<td>12-24G</td>
116-
<td>1</td>
117-
<td>2</td>
118-
<td>3</td>
119-
</tr>
120-
<tr>
121-
<td>Within 1000000</td>
122-
<td>24-48G</td>
123-
<td>1</td>
124-
<td>2</td>
125-
<td>3</td>
126-
</tr>
127-
<tr>
128-
<td>Within 2000000</td>
129-
<td>48-96G</td>
130-
<td>1</td>
131-
<td>2</td>
132-
<td>3</td>
133-
</tr>
134-
<tr>
135-
<td>Within 10000000</td>
136-
<td>128G</td>
137-
<td>1</td>
138-
<td>2</td>
139-
<td>Please contact Timecho Business for consultation</td>
140-
</tr>
141-
<tr>
142-
<td>Over 10000000</td>
143-
<td colspan="4">Please contact Timecho Business for consultation</td>
144-
</tr>
145-
</tbody>
146-
</table>
147-
148-
> Flexible memory configuration options are provided. Users can adjust them in the datanode-env file. For details and configuration guidelines, please refer to [datanode-env](../Reference/System-Config-Manual.md#_3-2-datanode-env-sh-bat)
149-
150-
151-
## 3. Storage (Disk)
152-
### 3.1 Storage space
153-
Calculation Formula:
154-
155-
```Plain
156-
Storage Space = Number of Measurement Points * Sampling Frequency (Hz) * Size of Each Data Point (Bytes, see the table below) * Storage Duration * Replication Factor / Compression Ratio
157-
```
158-
159-
Data Point Size Calculation Table:
160-
161-
<table style="text-align: center;">
162-
<tbody>
163-
<tr>
164-
<th>Data Type</th>
165-
<th>Timestamp (Bytes)</th>
166-
<th> Value (Bytes)</th>
167-
<th> Total Data Point Size (Bytes)
168-
</th>
169-
</tr>
170-
<tr>
171-
<td>Boolean</td>
172-
<td>8</td>
173-
<td>1</td>
174-
<td>9</td>
175-
</tr>
176-
<tr>
177-
<td> INT32 / FLOAT (Single Precision)</td>
178-
<td>8</td>
179-
<td>4</td>
180-
<td>12</td>
181-
</tr>
182-
<tr>
183-
<td>INT64 / DOUBLE (Double Precision)</td>
184-
<td>8</td>
185-
<td>8</td>
186-
<td>16</td>
187-
</tr>
188-
<tr>
189-
<td>TEXT (String)</td>
190-
<td>8</td>
191-
<td>Average = a</td>
192-
<td>8+a</td>
193-
</tr>
194-
</tbody>
195-
</table>
196-
Example:
197-
198-
- Scenario: 1,000 devices, 100 measurement points per device, i.e. 100,000 sequences in total. Data type is INT32. Sampling frequency is 1Hz (once per second). Storage duration is 1 year. Replication factor is 3.
199-
- Full Calculation:
200-
```Plain
201-
1,000 devices * 100 measurement points * 12 bytes per data point * 86,400 seconds per day * 365 days per year * 3 replicas / 10 compression ratio = 11 TB
202-
```
203-
- Simplified Calculation:
204-
```Plain
205-
1,000 * 100 * 12 * 86,400 * 365 * 3 / 10 = 11 TB
206-
```
207-
### 3.2 Storage Configuration
208-
209-
- For systems with > 10 million measurement points or high query loads, SSD is recommended.
210-
211-
## 4. Network (NIC)
212-
When the write throughput does not exceed 10 million points per second, a gigabit network card is required. When the write throughput exceeds 10 million points per second, a 10-gigabit network card is required.
213-
214-
| **Write** **Throughput** **(Data Points/Second)** | **NIC** **Speed** |
215-
| ------------------------------------------------- | -------------------- |
216-
| < 10 million | 1 Gbps (Gigabit) |
217-
| ≥ 10 million | 10 Gbps (10 Gigabit) |
218-
219-
## 5. Additional Notes
220-
221-
- IoTDB supports second-level cluster scaling . Data migration is not required when adding new nodes, so there is no need to worry about limited cluster capacity based on current data estimates. You can add new nodes to the cluster when scaling is needed in the future.

0 commit comments

Comments
 (0)