When computer storage and data management vendor NetApp built its second Global Dynamic Lab (GDL 2.0) at its Research Triangle Park, North Carolina-based, 155,000-square foot research and development facility, the company was looking for a solution that simplified data center operations. The Cisco® Application Centric Infrastructure (Cisco ACI™) did just that.
Al Lawlis, NetApp's senior director of engineering services, said the company assessed various solutions, but its strategy was to build on the next-generation Cisco Nexus® 9300 and 9500 switches—a key component of Cisco ACI, a solution that also features an Application Policy Infrastructure Controller (APIC).
GDL 2.0 is a 25 million watt facility that features almost 2300 racks of equipment and is powered by the Intel® Xeon® processor-based Cisco® Unified Computing System™, so getting the right solution for a facility of this scope was imperative. Cisco ACI fit the team’s needs, Lawlis says.
"Our Engineering Services organization acts as a service provider for NetApp’s engineering teams worldwide, and we intend to use ACI along with OpenStack to implement profile-based automation for our infrastructure," he says. “The advantage here is that teams will be able to select resources and services from a catalog, then click and deploy infrastructure-as-a-service in minutes or hours.”
The NetApp Engineering Services Team—one of the first adopters of the solution—will experience improvements in the density of throughput and cost per port, thanks to hardware upgrades enabled by Cisco ACI. The result: increased productivity of the infrastructure.
The architecture adds an extra layer of capabilities for the engineering team, which started deploying the solution in the spring of 2014, such as:
- A more stable testing environment
- Improved analysis of operations
- Better self-service
- Converged network management
Lawlis explains that the NetApp lab has always been set up as an entirely routable entity.
"With ACI policy management and service permissions, we will have integrated visibility and intelligence into what is happening in the network,” he says. “When we have application or tenant issues, we want to know the exact traffic information per tenant in the fabric.”
The Cisco ACI approach aligns with NetApp’s vision to make data centers more agile, secure, and responsive. “One of the basic tenets of NetApp’s storage operating system, clustered Data ONTAP, is that data is not bound to any single storage controller and is free to move around without disruption to the data access,” Lawlis says. “This approach is very consistent with the agility enablement of the entire application infrastructure stack provided by ACI.”
Running a tighter shop
Cisco ACI should save Lawlis’s team plenty of time. For instance, the architecture will help NetApp's recent initiative to free up the services team by adding user self-service tools. "ACI, with its clean interface and its policy-based orchestration, will enable the development of many more self-service capabilities," he says.
"We expect there will be significant increases in speed and productivity. ACI's policy and ACLs [access control lists] will create the environment for us to run a tighter shop."
Lawlis expects that application policy-based central services will make the triage and forensic investigation his team does much easier to understand while dramatically speeding up the change management process, because everything can be logically viewed from an application perspective.
Manage applications, not infrastructure
“By going above the infrastructure level right up to the application level, ACI is the ideal impetus to a converged network,” Lawlis says. He believes that managing a data center as a means for providing an environment for an application makes more sense than managing the network, storage, virtualization, and compute resources separately.
"People and organizations need to embrace the concept of managing applications through converged, multidisciplinary teams of engineers and admins,” he advises. "This will challenge many organizations that are organized along silos aligned to various infrastructure types, such as storage, virtualization, compute, network, and network services," he adds.
Lawlis has already consolidated his team, and advises potential Cisco ACI customers to consider following suit. Until those silos are broken down, he says, "organizations are not going to be able to really scale the way that ACI enables."