Intel launched the OpenStack Innovation Center to build new pools of developers who write applications that run on enterprise clouds, and who understand that as they write the code for these clouds, it’s based on best practices of operational experience.
Cloud Application Architecture
When application architects understand what’s happening at the infrastructure level, they want to be able to affect those decisions to improve performance. If I see that a workload is affecting another workload, or a high-priority workload (such as payroll,) I need to make sure that my applications get access to the resources they need. Intel’s Resource Director has the ability to tie the RDT features up through the layers in the infrastructure stack so it can benefit the application. Workloads can also be placed to take advantage of specific performance gains offered by certain servers, such as those offering an expanded CPU instruction set.
Intelligent Enterprise Data Centers: SNAP
How can enterprise customers use these features? Intel has been working on these features sets for years, providing visibility into into their platform. SNAP is an Intel-authored open source telemetry framework that provides detail about what’s happening on their platform into the hands of developers, so that those developers can do something with it.
Visibility is key to better decision-making in the data center.
One of the key things our team thought about when launching SNAP is that agents are difficult to operationalize because they have to be maintained individually. We asked, how can we reduce the operational burden of telemetry to zero?
Intel + Iron.io: Event-driven Job Processing for Anyone’s Data Center
Iron.io is leveraging SNAP as the way to understand the landscape of the data center, to better land workers on hardware that provides the most advantageous instruction sets, from crypto to FPGA to a vector instruction set accelerator.