About Host sFlow
The Host sFlow agent exports physical and virtual server performance metrics using the sFlow protocol. The agent provides scalable, multi-vendor, multi-OS performance monitoring with minimal impact on the systems being monitored.
Current approaches to server performance monitoring are highly fragmented. Each operating system, server vendor and application developer creates their own agents and software for performance monitoring, none of which interoperate.
In networking, standards and interoperability are extremely important and the problem of scalable, multi-vendor network performance monitoring has been solved. Most network equipment vendors support the sFlow standard and integrate performance monitoring in their products (see Network Equipment) allowing network managers to choose from a wide variety of open source and commercial network mangement applications (see sFlow Collectors).
Network convergence, virtualization and cloud computing blur the line between network and system management. For example, virtual machines are connected to the network through a software virtual switch. Monitoring the virtual switch requires an sFlow agent running on the server (see Open vSwitch). To understand performance in this environment you need to monitor both network and server resources and the sFlow standard includes physical and virtual server performance metrics.
The Host sFlow project provides an open source implementation of the sFlow standard, reporting on server performance. Computer vendors are encouraged to simplify deployment of sFlow by integrating the agent in their products. The combination of host performance data from servers and network performance data from switches provides a complete picture of performance in converged environments. sFlow provides a common standard for network and system management, reducing the number of tools needed to provide the unified, end to end view of performance required to manage converged, virtualized and cloud data centers.
In addition, the Host sFlow project forms the core of an ecosystem of related projects that extend visibility into the application layer; for example reporting web server response times, URLs, status codes etc.