Turbonomic Blog

Turbonomic & Cisco Tetration Are Enabling Network-Aware Automated Placement in Hybrid Cloud

Posted by Asena Hertz on Jun 26, 2017 8:30:50 AM
Asena Hertz
Find me on:

Cisco Live 2017 is here at last and we’re thrilled to announce our collaboration with Cisco Tetration Analytics to deliver the only network-aware self-managing platform for hybrid clouds. For a live demo, you can find us at the Turbonomic booth #4633. Enlin Xu, Sr. Engineering Manager at Turbonomic will be there to showcase the integration and answer all your questions.

The network touches everything: connecting people, devices, applications—and an exponentially increasing number of application components. Today, Gartner predicts that by 2020 there will be 20.4 billion connected devices, three times that of 2016. Capgemini predicts that by then cloud native application—apps and services built to perform optimally in the cloud—adoption will double. Containers and microservices are steadily becoming mainstream. And hybrid cloud is the preferred enterprise strategy: 85% of enterprises have a strategy to use multiple clouds with 58% planning on hybrid.

These trends are transforming IT from infrastructure-centric to application-centric, from hardware-centric to software-driven, and now hybrid cloud—bringing the criticality of the network to bear. The network has transformed from static matrices of cabling to global dynamically managed software-defined systems of connectivity. The question becomes how do you manage a complex, liquid supply of network resources? The answer is: “You don’t.” You let software do it.

 

Virtualization Brought Shared Resources, SDN Brought Liquidity

When it comes to the network there are tradeoffs between the fundamental laws of physics and the realities of shared virtual resources: 1) distance between connected things increases latency; 2) workloads localized to reduce latency contend for the same compute and storage.

With SDN the network becomes liquid and the tradeoffs between workload localization to reduce network latency vs. access to compute and storage have real implications for application performance. The network is more complex than ever. With more components that need to communicate, east-west traffic increases significantly and networks become more dynamic. It creates new challenges, including:

  • How do you minimize latency between endpoints?
  • How do you avoid top-of-the-hierarchy congestion?
  • How do you support “Tenant vMotion” across private and public clouds?
  • How do you accelerate network access to big-data?
  • How do you shape traffic at ingress ports?

 

The Economic Solution

Cisco Tetration Analytics provides a deep understanding of application dependencies and communication flows. This information provides valuable inputs to determine which workloads need to be localized. Communication flows, specifically, indicate how important that localization is versus the workload’s compute and storage requirements.

Turbonomic’s patented Economic Engine continuously analyzes the Tetration application dependencies and flow data, along with the compute and storage data from other sources (e.g., UCS Manager), matching the workload demands with supply on and off prem. It ensures that workloads get the resources they need when they need them, delivering real-time self-managing infrastructure that takes the network (and compute and storage) into account when making placement and scaling decisions.

 

What’s ahead?

This is just the first step in our joint effort to deliver our customers an autonomic system with network-aware placement. Together with Tetration, we also plan to add Application Dependency Mapping from the Tetration API. For customers that are adopting hybrid cloud, incorporating ADM will help ensure that customers are moving the right workloads together to the cloud.

 

To learn more and get a live demo, you can find us at the Turbonomic booth #4633. 

 

Topics: Cloud

Subscribe Here!

Recent Posts