How much to pay for protection in fiber access networks: Cost and reliability tradeoff
ABSTRACT The economical aspects are most critical in the access part of the networks because of the low sharing factor of the total cost. It makes the deployment of fiber access networks relatively slow due to the high investment cost involved. Several fiber access network architectures such as homerun (sometimes referred to as point-to-point P2P), active optical networks (AONs) and passive optical networks (PONs) have been developed to support the growing bandwidth demand from both residential and business customers. Meanwhile, fault management and reliable service delivery becomes more significant due to increased dependency on electronic services all over society and the growing importance of reliable service delivery. However, there is a tradeoff between the cost of protection and the level of service reliability. Obviously, in the access part of the networks, improving reliability performance by duplication of network resources (and capital expenditures CAPEX) can be too expensive. Therefore, a real challenge for an access network operator is to operate a cost-efficient, yet reliable network that provides leading edge services at low price and with minimal downtime. In this paper we present the evolution of fiber access networks and compare reliability performance in relation to investment and management cost for some representative cases. We consider both standard and novel architectures for deployment in both sparse and dense populated areas. While some recent works focused on PON protection schemes with reduced CAPEX the current and future effort is put on minimizing the operational expenditures (OPEX) during the access network lifetime.
- SourceAvailable from: Antonio Estepa
[Show abstract] [Hide abstract]
- "Network failures involve operational costs (OPEX) such as repair 7 and occasional penalties due to service interruption. These penalties can be due to external SLAs (in the case of operators ) or simply internal cost (in the case of a private LAN.) Nevertheless, the OPEX is not considered to be part of the RCND problem. "
ABSTRACT: Over the useful life of a LAN, network downtimes will have a negative impact on organizational productivity not included in current Network Topological Design (NTD) problems. We propose a new approach to LAN topological design that includes the impact of these productivity losses into the network design, minimizing not only the CAPEX but also the expected cost of unproductiveness attributable to network downtimes over a certain period of network operation.IEEE Communications Letters 03/2011; 15(3):349-351. DOI:10.1109/LCOMM.2011.012511.101742 · 1.46 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Industrial plants use conventional local area networks (LANs) to access a growing number of client/server (C/S) applications such as customer relationship management (CRM) or enterprise resource planning (ERP) which have a direct impact on organization’s productivity. These LANs are typically extended throughout the plant which makes them exposed to occasional accidents such as fiber breakages or power failures. Reliable network design (RND) problems address the design of minimum-cost topologies resilient to link failures up to a certain degree. However, RND problems fail to capture some parameters of practical importance for organizations such as productivity losses due to network outages, the time period for which the network design is expected to be operating, or the fact that not all nodes are equally important for productivity. We propose a new approach to LAN topological design named Productivity-aware reliable network design (PA-RND) that takes into account the productivity associated to each node of the network, minimizing not only the CAPEX but also the expected cost attributable to network downtimes over a certain period of network operation. Results show that our PA-RND problem optimizes the LAN topological design obtaining better results than current network design problems such as reliability constrained network design (RCND), minimum spanning tree (MST) or minimum cost ring (MCR).Computer Networks 06/2011; 55(9):2303-2314. DOI:10.1016/j.comnet.2011.03.011 · 1.28 Impact Factor