Cloud adoption is still the most vexing factor in increased network complexity, ahead of the internet of things (IoT), software-defined networking (SDN), and network functions virtualization (NFV), according to a new survey conducted by Kentik at Cisco Live 2017, Cisco's annual conference.
In addition, while machine learning is strongly embraced as an important technology for network management, most organizations aren't yet ready for network automation. In fact, most are still in the process of gaining sound operational visibility, integrating network management tool stacks, and implementing distributed denial-of-service (DDoS) security for their cloud and digital initiatives.
Key findings include:
Cloud adoption is still the largest factor in increased network complexity
36 percent of respondents indicated that the cloud adds the greatest network complexity to their organization, topping IoT (21 percent), SDN (12 percent), and NFV (3 percent).
Most organizations still have room to improve operational visibility for cloud and digital business networking
Only 20 percent of survey respondents think their organizations are doing an excellent job of monitoring the performance and security of their cloud and internet dependencies (e.g. IaaS, PaaS, SaaS, web APIs and web services). Another 25 percent reported that their organizations are doing a below-average to poor job.
Organizations need better DDoS detection capabilities
Despite the spike in DDoS attacks, including those hitting Dyn in October and Cloudflare in December, only 32 percent of respondents reported that their company is using DDoS detection technology to manage security of their cloud and internet dependencies.
Most organizations lag in integrating their management tool stacks
70 percent of respondents recognized that using the same stack of tools to manage both network performance and security can significantly improve operational efficiencies. However, the majority of respondents (59 percent) said their organization is not yet using the same stack of tools to manage both network performance and network security.
Machine learning is a priority, but most aren't ready for automation
60 percent of respondents said machine learning is "extremely important" or "very important" for network management. However, only 14 percent said their organization is ready for full network management automation.
"There is a lot of noise in our industry right now about intuitive systems and new-age machine learning that can monitor, identify and react to network conditions before issues occur. However, dozens of our largest customers have been telling us, and our survey results from Cisco Live support, that the key 2016 and 2017 enterprise efforts have focused on getting complete visibility into increasingly hybrid network complexity; detecting and preventing DDoS; and integrating tools that can provide operational and business value from network analytics," said Avi Freedman, co-founder and CEO of Kentik. "Full automation outside of constrained data center and cloud topologies is still a vision that customers are tracking, but network operators say that they need deeper and comprehensive visibility into their network's performance and security before they can let their networks run autonomously."
"Real-time network traffic intelligence is a critical component for network operators supporting their organizations with digital transformation," he added.
Survey Methodology: Kentik's findings are based on responses from 203 IT professionals surveyed during Cisco Live 2017. The respondents spanned more than 12 industries, including education, government, healthcare, finance, retail, software, telecommunications and transportation sectors. Respondents varied in job titles, from network engineers and network architects, to infrastructure managers, directors and executives. The majority of respondents came from organizations with 1,000 or more employees.
