The term server management conjures up different connotations in the mind of the listener. Depending on the type of server – software application server, virtual server or physical server, the issues they care about are different. Two tasks that instantly come to the fore are server configuration and server monitoring.
A software application server manager may visualize configuration of production middleware servers and the parameters may include database connections, memory size etc. A manager responsible for the virtual infrastructure in a data center may picture server configuration tasks as storing and accessing virtual images, operating system types etc. for the virtual machines. An infrastructure manager responsible for physical servers will take into consideration power, firmware and network configurations for the server.
Once the servers have been deployed, they need to be monitored. Servers can be monitored for basic availability, feedback, threshold crosses or outright failures. A software development manager could look for faults in the server logs with regard to database connections or code exceptions. A virtual machine manager may be looking for the noisy virtual machine (VM) that affects neighboring VMs. A physical server manager may monitor any hardware, power or network connection failures. After detecting any faults, server managers need to analyze the cause of the problem, determine corrective course of action and make necessary configuration changes to fix the problem.
Typically, initial server management is done through a GUI interface. This is very useful during development and for visual operational monitoring. Next, development and operations staff would need a command line interface and scripting capability to speed up server starts/stops and configuration parameter changes. This allows for rapid and scripted deployment of servers. The ultimate capability is when the server exposes an API and a software program can automatically configure and monitor the server based on predetermined policies. This allows closed loop management of servers. Imagine an ideal scenario where a server goes down and the management software detects the fault, determines the cause, alerts the operations center and automatically restarts the server with a safe configuration. Cisco UCS provides all the underlying capabilities needed to make this a reality in the physical server realm.
Cisco UCS Manager is the embedded device manager that runs in the Fabric Interconnect and provides the necessary management capabilities. The Cisco UCS Fabric interconnect switch provides low-latency, lossless 10 Gigabit Ethernet and Fibre Channel over Ethernet (FCoE). The UCS Manager provides configuration and monitoring capabilities for managing the physical servers and the UCS devices which provide network access. For an in-depth overview of the UCS Manager check out the UCS Manager Architecture document.
To simplify server management the UCS Manager provides a GUI, a command line interface (CLI) and an XML application-programming interface (API). Users can write their own custom applications to manage servers. I would argue that users can create a “Hardware as a Service” offer in quick order. And to top that, with the UCS Platform Emulator IT practitioners can even take the UCS Manager for a spin without procuring a single piece of UCS hardware – all at no cost. No wonder several independent software vendors have used the API to create value-added IT solutions such as service assurance for the UCS.
In this recent interview given by Cisco Cloud CTO, Lew Tucker, you will see how an API such as the UCS Manager XML API , is a critical component of Cisco Unified Management and data centers of the future.