The proposed method consists of three techniques; 1) quick recovery from the status being orphaned with the aid of cloud storage services, 2) proactive fetching of chunks from the CCDN, and 3) the reduction of the number of requests to the CCDN with the notion of frames.
During such a suspension, the orphaned peer should fetch missing chunks from the CCDN to keep the playback of the live stream.
Usually, requests to the storage service are more expensive than requests to the CCDN (see  and  for the example of cloud price in Amazon).
The key idea of the second technique is to allow several peers selected from each tree to conduct a proactive fetch of chunks from the CCDN so that the delayed chunks due to the shortage of P2P capacity does not occur.
Eliminated cloud peers rejoin their tree, while keeping to fetch chunks from the CCDN till they become a child of new parents.
The objective of the third technique is to reduce the number of requests issued to the CCDN with the notion of frames.
The increase of frame size F reduces the number of requests issued to the CCDN inverse proportionally.
As for the metrics for evaluation, we consider: 1) the amount of data fetched from the CCDN (Section 6.
Each peer starts playback after buffering the stream for 15 seconds, and it sends a request for a chunk to the CCDN when the remaining time before playing back the chunk becomes 2 seconds.
944) of downloaded chunks must be fetched from the CCDN, where 0.
At first, we evaluate the amount of data fetched from the CCDN.
This is due to the reduction of the number of requests handled by the CCDN.