@matigo it has to do with the edge locations. Because your connections between the edge and API would ideally be persistent, that overhead is cut down, and the client only has to negotiate the TCP and SSL between them and the closest edge location
@matigo it has to do with the edge locations. Because your connections between the edge and API would ideally be persistent, that overhead is cut down, and the client only has to negotiate the TCP and SSL between them and the closest edge location