CANedge2 Docs, Release FW 01.04.02
If multiple servers are available (such as regional cloud server endpoints), make sure to select a server
with low device-to-server latency.
0.7.3 Managing devices
This section covers how the CANedge interfaces to an S3 server and how to manage devices:
0.7.3.1 Buckets
Each CANedge device is associated with a specific S3
bucket
name. The device will (depending on the
configuration) e.g. push files and look for configuration updates within that specific bucket.
Note:
The device is not able to create a bucket. Buckets must be created on the server before the
device is connected
0.7.3.2 Device ID
Each CANedge device carries a globally unique 4-byte device ID (e.g.
A1B2C3D4
). The device ID is used
in the communication with S3 so that multiple devices connected to the same bucket do not conflict.
The root path of a given device is given by the Bucket name and the unique ID:
[BUCKET_NAME]/[DEVICE_ID]
0.7.3.3 Log files
The device uploads closed log files to the S3 server. When a log file has been successfully uploaded, it is
deleted on the SD-card. Log files pushed to the server are named according to the following pattern:
[BUCKET_NAME]/[DEVICE_ID]/[SESSION_NUMBER]/[SPLIT_NUMBER]-[EPOCH_TIME].[EXT]
•
EPOCH_TIME
: Epoch upload time in HEX (e.g.
2020/01/01 00:00:00
becomes
5E0BE100
)
•
EXT
: one of the possible file extensions supported by the device.
Note:
Upload of many small log files has a higher network overhead and will result in lower throughput.
Select a higher log file
split_size
(refer the
Section) for better throughput if the
network stability allows it. For reliable networks, aim for at least 10 MB files.
Note:
If the server-to-device confirmation that a log file has been uploaded successfully is lost (e.g.
due to network problems), the device will upload the log file again. This can result in multiple identical
log files with different
EPOCH_TIME
timestamps. Copies of the same log file (same
SESSION_NUMBER
and
SPLIT_NUMBER
) should be ignored or removed.
Note:
If log file time splitting is used (rather than file size based splitting), consider to use the
split_time_offset
to avoid that all devices upload new log files to the server at the same time. For
more information, refer the
Section.
0.7. Remote access
51
