michelangelus - Fotolia
In modern IT and business, much is made of the relationship between developers and IT operations. But the relationship between data center infrastructure engineers and IT operations is critical as well.
IT infrastructure and emerging technologies create new challenges for IT ops teams, as hyper-converged replaces large-scale infrastructures, data center network speeds increase, and spinning disks give way to more solid media.
These changes mean that IT ops teams and engineering teams must communicate about new areas of complexity and the way in which hardware functions. Ops must understand how these infrastructure hardware changes affect the application stack, and seek out training and teamwork among all IT groups.
Faster networking equals distributed storage
Higher-speed networking enables data centers to adopt more distributed infrastructure. Traditionally, high cost and complexity limited fiber optics in the data center to the storage area network (SAN) environment. As higher-speed Ethernet standards emerged, along with faster copper-based connections, it became more affordable to reach higher speeds. This spreads out functions typically reserved for the SAN or network-attached storage across the data center and takes advantage of multiple hosts or aggregated endpoints.
But this decentralization of IT infrastructure, while more cost effective from a Capex standpoint, adds endpoints with monitoring and management demands to the larger SAN environment. This can dramatically increase the amount of time and number of tools it takes to trace, identify and diagnose issues. While infrastructure and hardware costs decrease, support time windows increase. It's often more difficult to determine how this soft cost affects the business and budgets -- and it places additional strain on the role of ops teams supporting IT infrastructure and emerging technologies.
Converged hardware calls for new troubleshooting methods
Virtualization and hyper-converged infrastructure make each piece of IT infrastructure hardware that much more critical. The role of that equipment continues to expand, with localized storage and solid-state drives (SSD) keeping larger amounts of data closer to compute. This makes it even more critical for data center hardware to interoperate without issues and for troubleshooting between IT operations and engineering staff to be efficient and precise.
For IT operations teams, the compute for a virtual environment residing with storage blurs the lines between the two resources and makes it more difficult to troubleshoot problems. The solution most likely will be to adopt a single tool from the infrastructure vendor, instead of a collection of tools from multiple vendors that monitor utilization and performance across various systems. Vendor-native monitoring and management tools often are more detailed with a single focus; products designed to provide an overview of the entire IT infrastructure estate across diverse vendors' systems generally sacrifice some of these details that would enable faster troubleshooting.
SSD changes application behavior
SSDs are unavoidable on any list of emerging IT infrastructure technologies. SSDs provide a hardware-based performance boost to applications, wiping out I/O challenges without the developers having to recode. Without the traditional storage bottleneck, other pieces, such as the network or the application stack, become the performance chokepoint. This puts IT operations teams in a new situation: They must shift focus from storage and look for different and new kinds of performance issues. SSDs are an overall improvement to storage, but because they behave differently than spinning disk storage, operations teams will see new application behaviors. For example, SSD additions could cause unexpected CPU or memory spikes because of eliminated wait time for I/O, or saturation of storage frames trying to transport large amounts of data. They can even cause timing issues within the application itself now that the normalized wait time for I/O no longer exists.
While hyper-converged infrastructure and SSDs are not as disruptive to the data center as virtualization was, IT ops teams still must adjust to these changing times. IT infrastructure and emerging technologies continue to rapidly evolve, even in mature areas such as virtualization. It's tricky to keep up with policy and procedure in this constant state of change.
For ops teams to succeed amid these changes, it requires a focus on training and a solid relationship with infrastructure teams.
Data center storage possibilities improved with server SSDs
SSD is beginning a takeover from HDD in data centers
Consider how VMs will access SSD storage for best performance