在傳統廣告推播應用當中，廣告訂閱者(Subscribers)會直接從電信業者(或內容提供商)下載大量的廣告檔案，因此需要大量的頻寬資源。然而，在多個檔案同時下載的過程中，伺服器的上傳
頻寬將是效能瓶頸的主要原因，這可能會造成網路壅塞、傳送檔案不完整以及播放失敗的情況發
生。雖然同儕(Peer-to-Peer；P2P)網路傳輸技術可以解決網路壅塞的問題，但是過去的研究因為沒有把一些重要參數，如傳輸檔案到達的截止時間(Deadline)、檔案傳輸的優先權以及傳輸檔案的大小都考慮進去，因此無法保證檔案可以在截止時間前傳送完畢。本篇論文中，我們提出一個在廣告推播系統中具時效性的P2P封包排程技術， 可以根據截止時間， 來決定每一個檔案傳輸的優先權順序，並且基於目前的可用頻寬和剩餘傳送檔案的大小，每一個檔案傳輸連線所使用的頻寬也會利用拉格朗日乘數(Lagrange multiplier)法適當地分配。在我們的實驗當中， 相較於過去的P2P傳輸演算法，將以實際的例子來證明我們提出的方法，可以有效地提高在截止時間前檔案到達的比例， 幫助收發有時效性的檔案。

In traditional advertisement push/broadcast applications, a lot of advertisement files will be directly downloaded from telcom operators (or content providers) to subscribers, and more bandwidth resources are
needed. However, the server’s upload bandwidth might be the performance bottleneck when many files are downloaded at meanwhile. It results in network congestion, incompletely received files and playback failure
possibly. Although Peer-to-Peer (P2P) techniques tackled problems of network congestion, past researches can not still promise that files can be received by the expected arrival deadline because no important factors,
e.g., the constraint deadline, priority and file sizes, are all considered. In this paper, we propose prioritized time-constraint P2P packet scheduling for advertisement push/broadcastsystems. According to the currently available bandwidth, transmitted file sizes and the expected arrival deadline, each priority of file transmission can be determined, and then bandwidth resources will be allocated properly using the Lagrange multiplier method. In our experiments, we will give an example and prove that the ratio of the just-in-time file arrival will
increase with distinct control parameters, e.g., deadline away and file sizes.