The focus of the paper is on resource engineering for supporting Service Level Agreements (SLAs) in IP networks. We review some of the recent developments in Internet service architecture and protocol developments as well as some of the most important challenges still to be considered. An important performance metrics that must be considered in developing the new Internet is regarding the delay and the delay variation/jitter. Connected with this, a case-study is reported on end-to-end delay performane in a simple model for best-effort Internet where the bandwidth is not a commodity. The case-study is based on a client-server simulation model. The client and the server nodes are running applications like HTTP and SMTP, and they are placed in two sites separated by an Internet cloud. Performance issues of Short Range Dependence (SRD) and Long Range Dependence (LRD) traffic under different resource control regimes are compared. The results highlight the importance of the queueing delay at the WAN ingress point, which is due to the significant bandwidth differences that may exist, in this case, between the LAN and WAN link layers. The results also highlight the role TCP window size and Frame Relay Permanent Virtual Channel (FR PVC) control mechanisms play in the provision of delay performance for Internet services.