In order to support the cooperation and operation of multiple DGI systems over a certain geographical area, we need to design a communication network that can meet the communication requirements of the DGI systems. There are many design issues to be determined for such a network, including network size and topology, traffic policy, link capacity, protocol stacks, etc. These issues can be investigated using an emulated network (e.g., software simulation via OPENT or other network simulators), or a hardware-based network platform (e.g., subnets with switches or routers and links), as shown in Figs. 1 and 2, respectively.
On one hand, when a DGI program is running on a specific platform (e.g., the Mamba board loaded with Linux OS) over a LAN or simulated network, the program generates packet flows (or traffic) of commands or data measurement on the network. The traffic amount and patterns impact the local or entire network behavior significantly. On the other hand, many network performance characteristics can limit or impact the functions of DGI, including traffic type, packet delay, and packet loss ratio, etc. For example, the queueing problem causes the malfunction of the grouping function of DGIs due to the excessive delay in packet delivery. The data loss causes the missing of some important commands for the destination to perform certain DGI functions. The packet priority and various communication protocols may result in different network performances for different traffic types and offered loads. The mutual impact between the DGI traffic and the network performance has not been well investigated in networking research. In summary, we need to thoroughly understand and answer the fundamental issues in the design of the communication network that supports the DGI operation: in the viewpoint of communication network, what DGI wants and what the network can provide?
For these reasons, it is critical to investigate the network with different traffic and network characteristics, and thus find out the network requirements to support the DGI systems. For example, what is the maximum delay allowed in the packet delivery over a LAN or WAN to support multiple DGI applications successfully? For large-scale grouping, what is the maximum delay allowed for packet delivery over an IP network? What is the traffic policy that can be implemented in the common backbone network to support the DGI applications in practice?
We will investigate these issues by building a test-bed for the communication network that is designed for DGIs.
In order to support the cooperation and operation of multiple DGI systems over a certain geographical area, we need to design a communication network that can meet the communication requirements of the DGI systems. There are many design issues to be determined for such a network, including network size and topology, traffic policy, link capacity, protocol stacks, etc. These issues can be investigated using an emulated network (e.g., software simulation via OPENT or other network simulators), or a hardware-based network platform (e.g., subnets with switches or routers and links), as shown in Figs. 1 and 2, respectively.
On one hand, when a DGI program is running on a specific platform (e.g., the Mamba board loaded with Linux OS) over a LAN or simulated network, the program generates packet flows (or traffic) of commands or data measurement on the network. The traffic amount and patterns impact the local or entire network behavior significantly. On the other hand, many network performance characteristics can limit or impact the functions of DGI, including traffic type, packet delay, and packet loss ratio, etc. For example, the queueing problem causes the malfunction of the grouping function of DGIs due to the excessive delay in packet delivery. The data loss causes the missing of some important commands for the destination to perform certain DGI functions. The packet priority and various communication protocols may result in different network performances for different traffic types and offered loads. The mutual impact between the DGI traffic and the network performance has not been well investigated in networking research. In summary, we need to thoroughly understand and answer the fundamental issues in the design of the communication network that supports the DGI operation: in the viewpoint of communication network, what DGI wants and what the network can provide?
For these reasons, it is critical to investigate the network with different traffic and network characteristics, and thus find out the network requirements to support the DGI systems. For example, what is the maximum delay allowed in the packet delivery over a LAN or WAN to support multiple DGI applications successfully? For large-scale grouping, what is the maximum delay allowed for packet delivery over an IP network? What is the traffic policy that can be implemented in the common backbone network to support the DGI applications in practice?
We will investigate these issues by building a test-bed for the communication network that is designed for DGIs.