Updated AWS IoT (markdown)

s-hadinger 2019-06-10 16:45:14 +02:00
parent 53e25feb13
commit 5e93bd3e86
1 changed files with 20 additions and 2 deletions

@ -309,12 +309,30 @@ Memory consumption (nominal):
Note: if you use USE_WEBSERVER, your impact is lowered by 2k since the Web log buffer is reduced from 4k to 2k. Overall when activating USE_WEBSERVER, you just see a memory impact of 5.4k.
Memory needed during connection (TLS handshake):
Memory needed during connection (TLS handshake - fingerprint validation):
* ThunkStack = **4604 bytes**
* ThunkStack = **4608 bytes**
* DecoderContext = **1152 bytes**
* PrivateKey = **0** (stack if USE\_MQTT\_AWS\_IOT\_SKEY\_ON\_STACK) or 104 (heap)
* Certificate = **0** (stack if USE\_MQTT\_AWS\_IOT\_SKEY\_ON\_STACK) or 1040 (heap)
* Client Connect = **360 bytes**
* **Total for connection = 6.1k** (or 7.3k if private key forced on heap)
Memory needed during connection (TLS handshake - full CA validation):
* ThunkStack = **4608 bytes**
* CA Certificate in RAM = **328 bytes**
* DecoderContext = **3072 bytes**
* PrivateKey = **0** (stack if USE\_MQTT\_AWS\_IOT\_SKEY\_ON\_STACK) or 104 (heap)
* Certificate = **0** (stack if USE\_MQTT\_AWS\_IOT\_SKEY\_ON\_STACK) or 1040 (heap)
* Client Connect = **360 bytes**
* **Total for connection = 8.4k** (or 9.5k if private key forced on heap)
### Connection Time
ESP8266 is quite slow compared to modern processors when it comes to SSL handshake. Here are the observed performance to connect to an SSL/TLS server, depending on frequency (80MHz or 160MHz):
AWS IoT Connection, with EC Private Key
* **0.7s** at 160MHz
* **1.4s** at 80 MHz