As energy consumption has become a major concern of networks, traffic grooming studies on Internet protocol over wavelength division multiplexing (IP-over-WDM) networks have been revisited for energy-efficient optical network applications. Traffic grooming techniques can save on the number of installed lightpaths and corresponding IP router line cards at the cost of traffic processing overhead. Energy awareness in networking device development has initiated improvement of the energy-traffic proportionality of networking devices. Accordingly, networks will consist of networking devices with a wide range of energy-traffic proportionalities. Therefore, previously known traffic grooming techniques may not achieve energy-efficient networking in high energy-traffic proportional networks. Moreover, the traffic grooming technique tends to prefer longer paths, which are responsible for longer transmission delays. These two concerns trigger investigation of new aspects of heuristic traffic grooming algorithms for green and low delay optical networks consisting of network devices with diverse energy-traffic proportionalities. This paper demonstrates that a heuristic traffic grooming algorithm enhanced by comparison with direct bypass routing improves both energy and delay performances in a high energy-traffic proportionality regime. A hottest-first sorting policy in flow provisioning can further improve delay performance even in a low energy-traffic proportionality regime.