8. The gap analysis in terms of partial deployment and overhead seems to be lacked. A more complete discussion about the gaps would be helpful in understanding all the requirements mentioned here. From Yuanyuan Zhang [after ietf 114] #8
Response and revision summary: Existing SAV mechanisms do not have obvious gaps in partial deployment. Being efficient under partial deployment is only discussed in the requirement section as one of the necessary requirements for future mechanisms. High operational overhead is a gap faced by some of the existing mechanisms due to their technical limitations and has been presented in the draft. Other kinds of overhead like data-plane validation overhead is more like an implementation problem. So, it is not included in the requirement section (also following the suggestions of someone). But, we still put the descriptions “no packet modification” in the scope of SAVNET work, so that the data plane overhead can be controlled.
Response and revision summary: Existing SAV mechanisms do not have obvious gaps in partial deployment. Being efficient under partial deployment is only discussed in the requirement section as one of the necessary requirements for future mechanisms. High operational overhead is a gap faced by some of the existing mechanisms due to their technical limitations and has been presented in the draft. Other kinds of overhead like data-plane validation overhead is more like an implementation problem. So, it is not included in the requirement section (also following the suggestions of someone). But, we still put the descriptions “no packet modification” in the scope of SAVNET work, so that the data plane overhead can be controlled.