Measurement and Analysis of TCP Throughput Collapse in Cluster-based Storage Systems
Cluster-based and iSCSI-based storage systems rely on standard TCP/IP-over-Ethernet for client access to data. Unfortunately, when data is striped over multiple networked storage nodes, a client can experience a TCP throughput collapse that results in much lower read bandwidth than should be provided by the available network links. Conceptually, this problem arises because the client simultaneously reads fragments of a data block from multiple sources that together send enough data to overload the switch buffers on the client’s link. This paper analyzes this Incast problem, explores its sensitivity to various system parameters, and examines the effectiveness of alternative TCP- and Ethernet-level strategies in mitigating the TCP throughput collapse.

More on this white paper

 

The Case for the Unified Network Fabric
Historically, technology innovation began in the enterprise, where the IT organization would evaluate and slowly integrate it into corporate business processes. A subset of that technology would then make its way to consumers. Today, the cycle has largely reversed (Figure 1). Now technology innovation increasingly tends to occur in the home first, then makes its way into the enterprise via the employees, not the IT organization. Instant messenger, cell phones, Skype and other communication technologies were this first wave of the "consumerization" of the enterprise. These consumer technologies have been instrumental in increasing productivity by enabling workers to fully integrate their personal and professional lives.

More on this white paper

 

Open Network Automation is Critical to the Virtual Data Center

Mr. Kerravala is a distinguished research fellow at The Yankee Group, leads the Research Council and is chartered with the responsibility of providing thought leadership to the research organization. Comprising senior research leaders, the Research Council provides outreach to clients and the broader Yankee Group community, as well as ensures that the company's research agenda addresses the needs of business leaders. Kerravala drives the strategic thinking of the research organization and helps shape the research direction.

More on this white paper

 

Force10 Networks Dynamic Cloud Infrastructure

The data center has been, and continues to be, one of the key resources of innovation helping businesses become more agile and efficient while reducing total cost of ownership (TCO) — as well as being the focal point for green IT initiatives. The first steps are widely underway, with data center consolidation bringing back the idea of the centralized glass room, enabling better control over fixed costs. Virtualization helps drive up utilization of existing server compute and storage assets, reducing sprawl, while also helping to reduce the power and cooling footprint.

> More on this white paper

 

Distributed Core Architectures Using the Z9000 Core Switching System
As data centers scale to support thousands of servers, IT managers are seeking better ways to network those servers while reducing costs and power consumption. Moreover, in large-scale data center cluster environments inter-node communication bandwidth is increasingly becoming the main bottleneck. For these environments, applications need to exchange information with remote nodes for execution of their local computation. For example, web search engines require parallel communication with every node in the cluster to provide the most relevant results, and web servers may require interaction with hundreds of sub-services running on remote nodes.

More on this white paper