Using a cloud platform like Google's can enable network engineers to shift to a strategic focus.
As enterprises move their workloads to cloud providers such as Google, network engineers may wonder what role they will play in the future. Of course, they will continue to have a role in campus networking – running wireless access points, or WANs (in particular SD-WANs) to connect far-flung offices. But what about in the data center as their workloads move to the cloud?
Google offered a glimpse into the future for network engineers at the Google Cloud Next 2017 conference earlier this month. Google demonstrated how a large cloud service provider can help network engineers shift from the drudgery of maintaining the basic plumbing to a more strategic focus to help advance the business.
Google created a world-class network to run its workloads including its apps, search and advertisement infrastructure, and is now making it available to users of Google Cloud. Two elements are important: the physical infrastructure and the way the network is presented to end-user workloads.
Google built a fast, worldwide private network, so much of the traffic does not traverse the public internet. This provides better quality of service and user experience for workloads because the large number of direct connections enable fewer handoffs, lower latency, and less data loss. This type of global network is difficult for typical enterprises to replicate.
Networking as a Service
By using Google Cloud, the engineer doesn't need to worry about the global network or local configurations. The networking infrastructure is provided as a service to the server workloads, so the endpoints -- virtual machines, for example -- plug into the Google network, and all the back-end work to maintain and operate the network is performed by Google. Therefore, conventional work done by network engineers in configuring top-of-rack switches, setting up ADCs or load balancers is performed for them.
Google Cloud also provides simple access for network services and built-in software-defined networking. Typical networks require careful planning of configurations, service chaining, and security policy enforcement. This becomes extra challenging as engineers meld the behavior of physical devices with logically defined policies. Google Cloud leverages SDN, so the underling capabilities are easily exposed to enterprise users through user interfaces or APIs.
Does this mean there is no role for the network engineer in maintaining cloud provider networks like Google Cloud? Of course, there is work, but it is of a different sort. While the work to keep the network plumbing running and stopping any leaks is now handed off to Google, network engineers' responsibilities shift towards ensuring business requirements are met. Google’s network gives you a toolkit on which network engineers configure policies to meet users’ need. This still requires a good understanding of networking technologies, such as IP subranges, but engineers will spend less time dealing with details.
Moreover, since most enterprises continue to run their own private data center networks, they also need to help bridge traditional on-premises data centers with CSPs.
For a hybrid cloud environment that combines on-premises and off-premises resources, it will be necessary to harmonize the configuration, operational management, and monitoring of the cloud networks with the traditional on-premises network equipment.
Since it's impossible to install your own hardware within the Google infrastructure, it's necessary to look at gateways, virtual appliances, or cloud-based applications to create an integrated network, such as a mesh overlay network within the public cloud.
Integrating disparate systems is a skill that a good cloud-aware network engineer can bring. My recommendation for networking professionals is to understand how networking works in a CSP, ensure it meets business needs, and understand how to bridge that with existing networks.
Most importantly, spend more time with your line-of-business customers to understand the needs that drive the move to cloud computing, and make recommendations on how a powerful CSP network can work in concert with the existing network infrastructure. By making recommendations on how to meet business needs by making the best use of the current and future infrastructure, engineers will be valued by the organization, and prepared for an evolving cloud-centric world.
You can meet Dan Conde at Interop ITX in Las Vegas, where he will co-present "Things to Know Before You (Hyper) Converge Your Infrastructure." Register now for Interop ITX, May 15-19.