Industry executives and experts share their predictions for 2024. Read them in this 16th annual VMblog.com series exclusive.
AIOps Will Pervade the Data Center in 2024
By Ben Baker, Senior Director, Cloud/Data Center Marketing at Juniper Networks
Enterprise data centers face tremendous pressure to accommodate the explosive growth in data needed for modern digital businesses to operate, especially as they rapidly experiment with new AI applications.
In both scale and complexity, environments are pushing the boundaries of legacy data center architectures and networking. To be sure, compute and storage are strained as well, but these resources are readily available in the cloud and co-location sites. Networking is what keeps information flowing, serving as the lifeblood of all enterprise IT operations.
Network TLC
But networking is not a static resource. It requires care and feeding to ensure it can provide not just adequate but optimal performance. This encompasses everything from throughput and bandwidth flexibility to resource optimization and security, all of which are growing more complicated by the day.
The development and deployment of applications is integral to these efforts, and with the demand for new services accelerating at such a rapid pace, many organizations have implemented DevOps models of continuous integration/continuous deployment to ensure operations do not lag behind user demands. But even this is no longer fast enough for today’s highly digitized environments, which is why new forms of intelligence-driven AIOps are starting to make their way into network operations.
At the moment, most network facing AIOps workflows are centered on security and basic use cases within campus and branch environments. Looking ahead to 2024, this will shift to more operational aspects of networking, namely predictive maintenance and troubleshooting and additional domains including the data center.
Perhaps the singular advantage that artificial intelligence brings to network operations is the ability to analyze large amounts of data and then make recommendations for future action, or even perform those actions autonomously. This is a godsend to the networking team, which has been struggling to keep up with the triple demands of increasing complexity, skills shortages, and tight budgets.
Eyes on the Network
In troubleshooting, for example, AIOps can actively perform multiple steps in root cause analysis currently executed by people, then present the results in a cohesive, easily digestible manner. This significantly improves key metrics like mean-time-to-repair and mean-time-to-innocence, when app teams and networking teams are pointing the finger at each other. In addition, organizations are likely to accelerate the deployment of large language models (LLMs) across all interfaces on the network, providing ready access to documentation and performance data in order to assess current operating conditions and guide configuration changes and upgrades.
Equally important is the ability to actively monitor and optimize network systems and functions to maintain peak performance on a continual basis. This “self-driving network” results from AIOps’ rapid action cycle, which not only corrects problems as they arise but proactively manages all facets of network operations to prevent small anomalies from turning into larger issues that impair the user experience.
For instance, a sudden spike in network traffic can lead to bottlenecks and disruption of services. While this can be handled by a human once the alarm has sounded and the manual processes are put into motion, AIOps has proven itself to be far more adept at managing loads in real or near-real time in many situations. By constantly monitoring traffic flows, AIOps can proactively provision the necessary resources to accommodate spikes, and then just as easily restore the network to its prior state once the load has returned to normal – all without human intervention or noticeable degradation of service.
In 2024, we can expect to see the enterprise take full advantage of AIOps throughout the entire network operations and management stack. In a world where end users have zero tolerance for lagging application performance and outright downtime, the need to keep all systems in a constant state of availability and running at peak efficiency is paramount.
And best of all, this comes with a significant reduction in the overall cost of networking at a time when profit margins across all business models are becoming razor thin. AIOps will free up humans to work on more strategic, digital transformation projects that many of their CIOs are currently wrestling with.
##
ABOUT THE AUTHOR
Ben Baker is a Senior Director, Cloud and Data Center Marketing & Business Analysis at Juniper Networks. Ben has been working with service providers around the world for the last 20 years developing business cases for a variety of product concepts and new ventures. Ben holds an MBA from MIT and a BS & MS in Mechanical Engineering from Johns Hopkins University.
