Network Protection with Multiple Availability Guarantees
Abstract
We develop a novel network protection scheme that provides guarantees on the time a flow has full connectivity, and guarantees a quantifiable minimum grade of service during a downtime. In particular, a flow can be below the full demand for at most a maximum fraction of time; and then, it must still support at least a fraction q of the full demand. This is in contrast to current protection schemes that offer either full protection or availability-guarantees with no connectivity during the downtime. We develop algorithms for the single and multiple commodity cases for general networks, and show that significant capacity savings can be achieved as compared to full protection. For example, if a connection is allowed to drop to 50% of its bandwidth for 1 out of every 20 failures, then a 24% reduction in spare capacity can be achieved over traditional full protection schemes. For the case of q = 0, which is the standard protection constraint, an optimal pseudo-polynomial timed algorithm is presented.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 05, 2011
- Accession Number
- ADA573569
Entities
People
- Aradhana Narula-tam
- Eytan Modiano
- Greg Kuperman
Organizations
- Massachusetts Institute of Technology