一個針對行動無基礎網路中的信任證據產生、散佈與信任值計算機制

Abstract

身分認證在行動無基礎網路中是網路使用者識別的基礎方式，讓周圍節點確認往後將要通訊的對象。又為了衡量通訊節點的行為，可將節點識別和信任程度相互結合，以達到資料轉送的安全。然而在行動無基礎網路中的節點為互相獨立且為完全自我組成架構，所以必須擬定一套包含身分識別、憑證資訊、以及信任機制的整體信任證據運作，並在線上做信任證據產生、散佈、衡量、以及驗證。 本研究針對行動無基礎網路提出一個分散式的信任證據運作機制，讓網路節點可以各自產生代表己方的憑證並擁有相關的信任識別，而不用中央管理端控管，以作為節點的識別、傳送、衡量。信任證據管理方法可讓信任證據在網路中經由傳送來得到節點的資訊，且不會被假造或修改。節點在經過互動後，信任值高的節點可透過信任值重新衡量適時反應先前階段的互動結果。於實驗中證明此信任證據運作方式可以解決自私節點和惡意節點的問題，並可以在行動無基礎網路中得到最好的路由參考。運用賽局理論也可證明節點合作可以得到最大的利益，若中繼節點採取背離的策略，則此節點會因為信任值的降低而成為不受信任的節點，故傳送資料時也會遭受其他節點的拒絕，直到合作為止。

Personal identity is the basic way to present user’s role for MANETs, so all around nodes can verify the node which they will communicate with in the future. In order to evaluate the node’s behavior, we can combine the identity with trust value. But nodes are in independent and self-configured architecture, so it is important to develop a totally trust evidence which contains personal identity, certificate information, and trust operation mechanism. Moreover, the trust evidence can be established, distributed, evaluated, and verified on-line. This research proposes the distributed trust evidence operation mechanism for MANETs. The node establishes certificate itself and has corresponding trust identity without central certificate authority. The way to manage trust evidence can get others’ trust evidence via the transmission of packets and would not be modified by malicious nodes. The model will resolve selfish node and malicious node problem via simulation. It will be suitable to operate MANETs and provide most correct routing reference. We can also prove that nodes in MANETs will cooperate via game theory. After interaction, higher trust nodes can reflect the outcome and re-evaluate the trust. If the intermediate node deviates, its trust value will be decreased and will be regarded as the doubtful node. Therefore when the doubtful node requests surrounding node to forward packets, it will be rejected until the node cooperates.