CAR: Cloud-Assisted Routing

Case ID:
UNR13-008

Background

Multi-homing and peering practices, along with the demand for increased routing flexibility, have significantly contributed to the growing complexity in router operations. Operators now expect more programmability from routers, which further challenges traditional router architectures by necessitating the inclusion of software-based designs and the virtualization of routing as a service. Due to these evolving complexities, there is a pressing need for innovative techniques and architectural approaches that can reduce or offload these complexities from the routers.

 

Description

Researchers at the University of Nevada, Reno have developed an innovative architectural approach called Cloud-Assisted Routing (CAR). This approach leverages the high computational power and memory capacity of cloud services to simplify complex routing functions such as forwarding and flow-level policy management. A key feature of CAR is its balanced approach, which finds a middle ground between purely local scaling of router performance and entirely cloud-based, flexible routing services. The CAR technology offers significant advantages to the industry, as outlined below.

 

Advantages

  • Hybrid Dependency: CAR's design, which relies partially on cloud resources, allows routers to quickly reassume control of delegated tasks in the event of cloud-related connectivity or performance issues.
  • Cost Efficiency: By shifting complex tasks from hardware to the cloud, CAR enables the commoditization of router hardware, fostering a transition towards a software-oriented routing market. 
  • Cost-Effective Integration: The approach utilizes lower-cost commodity cloud services to supplement expensive router hardware, optimizing peering decisions by identifying the best routers to connect with.
    • Resiliency to Failures: Cloud-based forwarding and rerouting schemes improve network robustness.
    • Efficiency: Centralized cloud-based optimizations enhance intra-domain traffic engineering.
    • Economic Competitiveness: Cloud-based on-demand service provisioning enhances cost efficiency and flexibility.

 

Related Documents

 

Patent

 

 

UNR13-008

Patent Information:
For Information, Contact:
Ray Siripirom
Senior Licensing Associate
University of Nevada, Reno
csiripirom@unr.edu
Inventors:
Murat Yuksel
Tarik Karaoglu
Keywords:
Software