|
Abstract: |
Random contentions occur in optical burst-switched (OBS) networks because of one-way signaling
and lack of optical buffers. These contentions can occur at low loads and are not necessarily
an indication of congestion. The loss caused by them, however, causes TCP at the transport layer to
reduce its send rate drastically, which is unnecessary and reduces overall performance. In this paper
we propose forward segment redundancy (FSR), which is a proactive technique to prevent data loss
during random contentions in the optical core. With FSR, redundant TCP segments are appended to each
burst at the edge and modified burst segmentation is implemented in the core so that when a contention occurs,
primarily redundant data is dropped. We develop an analytical model of FSR and perform extensive simulations.
FSR is found to improve TCP's performance by an order of magnitude at high loads
and by over two times at lower loads. We also compare FSR to controlled
burst retransmission and find that FSR outperforms retransmission at the OBS layer.
We also find that FSR is more fair to background traffic than both burst retransmission and traditional OBS. |
