Server Management – Critical consideration for data center automation and orchestration
I wish the topic of Server management was as juicy as Lance Armstrong’s confession or as intriguing as Manti Te’os girlfriend hoax, but it is NOT. It is, however, intertwined with two of the domains– 1) infrastructure management and 2) automation and orchestration described in the Cisco Domain Ten model. Server management was also one of the topics in a series of discussions with Mike Spanbauer from Current Analysis.
Cisco UCS Management addresses the problems of complexity and scaling which the panel discussed. Service profiles ease the deployment of policy based server management and simplify routine setup tasks as well as server repurposing tasks within the typical server lifecycle. The Cisco UCS is architected for automation from the very core with an open API. The operational benefits can usually be quantified in $$ and cents as some of these customers have experienced.
Travelport – “86 percent savings in total support hours”
Hendrick Automotive Group – “Boost IT staff productivity by more than 30 percent”
Xerox – “Staff productivity has improved 20 percent…”
Peak10 – “We can certainly quantify the impact Cisco UCS has had on our operations: the benefits are clearly there…”
To create an effective computing platform that supports mission-critical applications and cloud-computing environments, automation and efficiency are basic requirements. With traditional rack and blade servers, the operational complexity of managing discrete infrastructure (server, networking, and storage access resources) with no automation, standards, or process repeatability, can pose considerable hurdles. Orchestration software is typically layered over the infrastructure, creating an accidental management architecture. These management tools can be rigid, and in the long term, increase costs through complexity and reduced efficiency. IT departments should therefore choose open infrastructure systems, which can be managed flexibly. This is even more important with the introduction of virtualized machines. Now bare metal servers and virtualized servers have to co-exist and managed with stiff service level agreements.
Automation and scripting capabilities improve repeatability and reduce process variability, which are important for any world-class industrial operation. These are principles we can learn from our manufacturing counterparts. Standards and process repeatability are sometimes underestimated, but when adopted, improve the overall quality of service delivered. The result of standardization and automation of server management is succinctly captured in the following excerpt from the Xerox Cisco UCS case study:
The IT team handling Xerox Cloud Services also saw efficiency improvements through the enhanced provisioning process. “With the integration and reliability from Cisco UCS in Vblock, we don’t have to worry about supporting different models. We just need to write one profile for all virtual servers, and then we can completely automate the rest of the provisioning process,” says Houser (CTO for Xerox Cloud Services). “This saves our team a tremendous amount of time and enables us to grow quickly.”