HP A7446B Converged fabrics: Emerging technologies for simplifying data center - Page 2
Abstract, Introduction, Traditional data center topology - target
UPC - 882780355118
View all HP A7446B manuals
Add to My Manuals
Save this manual to your list of manuals |
Page 2 highlights
Abstract Converged fabrics is a solution for simplifying data center management by consolidating all communication (server, storage, network, and remote management) onto a single fabric, in contrast to the more complex and costly approach of using proprietary fabrics for inter-process communication (IPC) and storage. This technology brief explains the emerging and growing unified fabric technologies in which HP is involved. These technologies include 10 Gigabit Ethernet (10GbE) and Fibre Channel over Ethernet (FCoE) and how they work together to tie multiple network fabrics into a single, converged infrastructure. Introduction The complications of implementation, management, and cost of multiple network fabrics have prompted HP and other vendors to investigate converged fabric solutions, also known as unified fabric or fabric consolidation. Any network topology constructed with one or more switched network nodes can be described as a fabric. Therefore, fabric is the term commonly used to refer to individual network types, including communication, storage, management, and high speed networks. Converged fabrics is a solution for simplifying data center management and reducing infrastructure costs. It consolidates all individual network types onto a single fabric, in contrast to the more complex and costly approach of using proprietary fabrics for IPC and storage. The converged fabrics solution is the common interconnect and can be either Infiniband (IB) or 10GbE. HP expects converged fabrics technology to significantly lower infrastructure and management costs without sacrificing performance. This technology brief focuses on how 10GbE will be used as the common interconnect for converged fabrics. Traditional data center topology Some common characteristics of current data center infrastructure are underused capacity, inflexible single-purpose resources, and costly management. The reason is that data center topology typically includes separate, heterogeneous network appliances to manage different types of data. Each of these appliances adds to the complexity, cost, and management of the data center. Multiple networks require unique switches, network adapters for each server, and network management systems targeted for each network. Businesses may support as many as four unique networks to manage Storage Area Network (SAN) data, IPC clustering data, remote management data, and Ethernet communications data. Each data type typically uses a different network and interconnects, which adds to the complexity of network architectures and to unifying network fabrics. 2