由於地理人口分佈的不均和網路接入點的存在使得網路流量請求與傳輸呈現區域性的集中，因此使用衛星星座傳輸的流量將會呈現地域性的差異，而這會導致衛星網路裡部分衛星鏈路擁塞或未得充分利用的情形。在隨著各項網路服務的推展，流量需求持續增長的背景下，為了保證衛星網絡在吞吐量和延遲有更好的性能，本文提出一套基於緩存移轉狀態的動態平衡路由協議，使鄰近衛星能夠交換有關其鏈路緩存隊列的狀態，讓發送數據的衛星能夠衡量各鏈路的和適度，流量在衛星之間更好的分配，減少發送數據衛星與相鄰衛星因擁塞所導致的數據包丟失。為了評估路由算法的性能，本文的研究建立了一套基於 Python 的模擬環境，模擬在理想的 Iridium-like 星座系統下，不同演算法在衛星節點中傳輸的表現，根據實驗結果，本文的演算法在吞吐量、丟包率和流量分配相較於靜態路由、TLR 和 QSDR 演算法，取得了更好的表現。 ;Due to uneven geographical population distribution, the presence of network access points, network traffic requests, and transmissions exhibit regional concentration. Therefore, the use of Satellite constellation transmissions will result in regional differences in traffic, leading to congestion in some satellite links or underutilization. With the continuous growth of traffic demand driven by various network services, to ensure better performance in terms of throughput and latency for satellite networks, this study in this thesis proposes a dynamic balancing routing protocol based on the transfer state. This protocol enables neighboring satellites to exchange information about their link queue states and allows sending satellites to evaluate and adaptively distribute data among the links, thereby reducing data packet loss caused by congestion between sending satellites and adjacent satellites. To evaluate the performance of the routing algorithm, this study establishes a simulation environment based on Python, where the performance of different algorithms in satellite node transmissions is examined in an ideal Iridium-like constellation system. According to experimental results, the proposed algorithm outperforms static routing, TLR, and QSDR algorithms in terms of throughput, packet loss rate, and traffic distribution.
