NetEx takes several innovative approaches in order to meet the objectives set forth in the previous section. These are in a way the guidelines by which the different components of NetEx are built.

• Complementation

NetEx is a middleware solution that provides delay-guaranteed communications service. It is designed to run on commercial host operating systems (like UNIX) and network platforms (like FDDI, ATM). The modular design of NetEx allows it to be configured to render functions that are necessary but complementary to those already provided by applications and/or operating system. It also allows applications to choose required NetEx functionality from a spectrum of host and traffic management functions offered. Consider the case of an app lication running over a real-time operating system that is capable of scheduling the host system resources. While NetEx has the gear in place to manage both host and network resources, it can be configured to just manage the network resources in this case. NetEx thus does not add any redundant services. The use of relevant existing facilities reduces the time taken for an admission decision. Thus, this complementation approach increases the system efficiency without sacrificing its effectiveness.
• Co-operative Management

The success of providing a premium communication service over a distributed system depends crucially on how the host and network resource managers co-operate. NetEx realizes co-operative management in several ways:
• NetEx appropriately partitions local and system wide management functions in its HTM and NTM
• NetEx gathers and uses information across the protocol stack layers: from the network layer to the applications themselves.
• NetEx allows applications to participate in the connection management and resource allocation by specifying directives that are passed on to the HTM and then to the NTM.
This co-operative management approach makes it possible for NetEx to offer best possible QoS to real-time connections. Such coordination ensures correctness, efficiency, and effectiveness in connection admission and run-time traffic control . As a result, NetEx easily scales with an increase in the number of nodes.
• Adaptation

Traditional Connection Management (TCM) does not exploit dynamic fluctuations in resource availability. This leads to gross under-utilization of resources. NetEx aims at maximizing resource utilization while providing maximum Quality of Service (QoS) possible with the available resources. NetEx achieves this improvement by providing Adaptive Connection Management (ACM). In the connection admission, ACM searches in the admissible QoS space in order to admit the new connection while maximizing the QoS provided to the new and existing connections. Furthermore, when a connection is torn down, ACM expands the offered QoS of existing connections to utilize the resources released by the outgoing connection. Thus, adaptation maximizes resource utilization and offered QoS.
NetEx also provides user control over the adaptation process. If adaptation has to be invoked while admitting or tearing down a connection, then the system follows the directives given by the user. NetEx can thus be sensitive to and consistent with the mission-specific requirements of the user.
• Traffic Modeling

Traffic modeling has a direct effect on the delay analysis, which is critical in offering delay-guaranteed communication services. Furthermore, proper traffic modeling is pivotal to the effectiveness and efficiency of the system. In NetEx, we use a multi-segment approximation method to describe traffic. While this gives a close representation of the real traffic, it is also simple enough to perform the delay computation efficiently.
• Closed-form Delay Analysis

NetEx is designed to provide delay guaranteed communication service. Before a connection can be admitted with a guaranteed delay, it has to go through delay computation for the CAC to decide admissibility. Many existing solutions on delay computation are numerical procedures that are typically time-consuming. Based on delay calculus, we have developed a closed-form formula for computing delays. There are several advantages: the closed-form delay analysis is much more efficient than numerical procedures. The delay bounds derived from the closed-form formula are relatively tight, hence improving the system effectiveness as well. Obviously, the simplicity of the close-form delay analysis helps NetEx to scale with the growth of the system in terms of both number of connections and the size of the network.