LwM2M Firmware repository - fragment with the MO itself to configure individual FW updates to use one protocol over another
Orignial question: We currently have support for both CoAP and HTTP based protocols for the firmware repo, how do we configure individual firmware updates to use one protocol over another?
Attached is further commentary between Elias and Cris Danci.
Thanks a lot for the feedback. We have been discussing this internally.
In the attached e-mail there was a question of whether FTP could be supported for firmware update image delivery. Unfortunately, this delivery mechanism is not part of the OMA LWM2M standard and hence there is no uniform way to support FTP delivery in a standard way across different device types. For this reason, we would only implement such a feature if this is specified by the OMA LWM2M specification.
There was an other question regarding changing the preffered delivery mechanism (pull vs push based) on a device level. Our implementation adapts to the device capabilities and selects the best mechanism. We believe that changing this behaviour is a possible source of error and added complexity. We have not yet seen a real-life case where this would really be required from a business perspective. Before doing any change in our implementation, we would like to understand a real scenario where this is a hard requirement.
Thanks a lot for the feedback. We have been discussing this internally.
In the attached e-mail there was a question of whether FTP could be supported for firmware update image delivery. Unfortunately, this delivery mechanism is not part of the OMA LWM2M standard and hence there is no uniform way to support FTP delivery in a standard way across different device types. For this reason, we would only implement such a feature if this is specified by the OMA LWM2M specification.
There was an other question regarding changing the preffered delivery mechanism (pull vs push based) on a device level. Our implementation adapts to the device capabilities and selects the best mechanism. We believe that changing this behaviour is a possible source of error and added complexity. We have not yet seen a real-life case where this would really be required from a business perspective. Before doing any change in our implementation, we would like to understand a real scenario where this is a hard requirement.