Research On Correctness And Optimization Technology Of LFB Topology In Forces System

The architecture of traditional router is closed and rigid. When new types of business need to be deployed in the traditional network, it requires a large range of equipment replacement and reconfiguration of complex network strategy, which will lead a very long cycle, huge workload and to be error-prone. Therefore, a novel network device with data plane resource open and programmable is much needed to solve the "rigid" problem of current network fundamentally. Based on this background, IETF ForCES technology came into being.ForCES is a technical system, including FE model and a specification of communication protocol between CE and FE. FE model is actually the essence of ForCES, it carries out the modular processing of FE which possesses data forwarding function. FE is abstracted into the organic combination of a number of LFB. These LFB are connected to each other, CE can control and define the connection between LFB (i.e. LFB topology) through standardized ForCES protocol. Different LFB topologies are ultimately reflected in different functions of FE (IPv4Forwarding, IPv6Forwarding, QoS, etc.).In addition, FE model also defines a special LFB (FEObject), this kind of LFB is not involved in packet processing. It works like an agent, who collects, saves and informs CE the current important information of FE, and the description of a LFB topology is one of the important content. The research on correctness guarantee mechanism and optimization technology of ForCES LFB topology in this article is expanded around FEObeject.Referring to the safeguard method of correctness of topology in OpenFlow, Click and data center network. We put forward a feasible safeguard mechanism of correctness of LFB topology in ForCES system. Firstly, read the ability parts’information of LFB whose class named FEObject, enumerate all the possible links between the LFBs to construct a topology ability link list. According to the target topology, construct all the actual links. Then use a string to represent a LFB link after the character processing. Finally, based on AC algorithm, an improved multi-pattern matching algorithm is proposed. Take all the actual links as pattern strings and put them into the LFB topology ability link list to match. So we can determine whether the actual LFB topology is correct or not through the matching result.Knowing the correctness of the LFB topology, the next step is optimization process. After detailed analysis, we learn that the LFB topology optimization problem is a NP-Complete problem. So we propose a new LFB topology optimization model on the basis of QoS guarantee. Under the premise of meeting the basic function the user requirements, QoS assurance, and minimum total cost of all links, then eventually finding a transmission route with the minimum of resource consumption, variance of link utilization, risk of path loss to transmit business.