Speed, simplicity, security and convenience are critical in a customer-first world, where one minute of downtime can mean long-term consequences for a business’s brand reputation and bottom line. When customer experience is everything, the underlying network infrastructure is essential for any business’s long-term success. Yet today, networks are more difficult than ever to build and maintain.
As recently as five years ago, networks represented sets of static links. IT teams and network administrators focused on connecting systems to private datacentres, manually ensuring point A could share information with point B. Networking looked far different from how it does today.
Fast forward to the present, and with the ubiquity of smartphones and unprecedented expectations for speed and convenience at scale, we have seen a sea-change in the way systems and applications are built. Businesses across industries have begun to embrace microservices, breaking up monolithic applications into multiple self-contained functions.
Although microservice architectures are intended to facilitate agility, they introduce new challenges. For an application to run seamlessly, all the microservices that comprise it must communicate with each other in a secure, performant, low-latency manner.
Because microservices have no limit in quantity and are tied together by networking technology, the scope and importance of the network has never been larger. Networking has become dynamic.
In turn, understanding and maintaining system health has become a herculean task that traditional IT point solutions simply cannot handle. And hybrid and multicloud environments generate even more heterogeneity and complexity, worsening the headache.
Openness, intrinsic security and proactive operations are new networking approaches. If the traditional IT toolset no longer cuts it, we need to build modern networks atop an entirely new set of technologies and frameworks. What are these?
1. Open standards
The notion of open network standards has gained renewed significance in the microservices-based era. Open application programming interfaces (APIs), by stitching together applications across on-premise and cloud, enable enterprises to preserve existing infrastructure investments while modernising at their own pace.
And open source software such as Envoy, Istio and Kubernetes allow standards to be developed transparently and through code, saving businesses from having to write APIs and models from scratch.
2. Dynamic networks
Second, dynamic networks have radical implications for security, and vice versa. While legacy networks only had a single ingress and egress point, network security today involves safeguarding an endless number of disaggregated endpoints.
This represents an opportunity to build security intrinsically into the network, rather than layering firewalls on top, to create a native, always-on security platform wherever the network lives. Because the network sees all traffic flows, it can detect when activity deviates from the norm and identify risks of attack.
3. Machine learning
Lastly, machine learning creates opportunities to tame dynamic networks. Until recently, network operations have been largely reactive. Misconfigurations were not discovered until after they were pushed to production, and network management teams spent most of their time simply troubleshooting problems.
But with machine learning-based technologies, businesses can, for example, monitor performance more quickly and at a tremendously larger scale than was previously possible. Machine learning tools can also help to prevent outages by proactively predicting the impact of a migrated workload or configuration change before it is deployed, and even recommend a fix.
As customer expectations inevitably continue to grow, so will network complexity. Networking is the critical backbone that makes sophisticated customer experiences possible, from premium video streaming, to massive multiplayer games, to seamless collaboration between colleagues who live halfway across the world, and more.
Those experiences will change as tastes and trends shift over time, and networking technology is constantly evolving, but one thing is for certain – the network matters more than ever.
Shailesh Shukla is vice-president and general manager for networking at Google Cloud.
Read more about networking trends
- As more organisations embrace DevOps in networking, the approach can transform network requirements, such as visibility and security policies, which networks teams should consider.
- Ansible is one of the most popular open source network automation tools. Network pros can use it to perform basic network automation tasks, such as collecting ARP table data.