Automating networks for whatever comes next [Q&A]

remote working

Digital transformation and modernization of IT is sweeping across many organizations at the moment. But one aspect that's sometimes neglected is their impact on networks.

How can enterprises scale their networks to cope with change and what part can automation play in the mix? We spoke to Ernest Lefner, chief product officer at Gluware, and co-founder and former co-chairman of ONUG (Open Network User Group), to find out.

BN: What are the IT trends impacting networking right now and how do you expect them to evolve in the coming years?


EL: Networks remain in the midst of a radical evolution that began in earnest with pandemic remote work measures. Enterprises initially reacted by rushing their network operations into the cloud using cloud native tools. We now see organizations stepping back in an effort to 'rightsize' their cloud use by more thoroughly assessing how many providers they want to utilize, what goes into the cloud, what stays on-premises, and more.

Cloud transformation and the rebalancing of networks have led many to believe that the number of network components like switches and routers will decrease over time. In reality, this number will continue rising exponentially as we ask more and more of our networks. This is a primary contributor to the next trend I see developing: Automation. We will soon reach a point of no return where the complexity of networks makes them impossible to manage without intelligent automation.

Lastly, there will always be a new IT security trend dominating the conversation. However, the core components of good security haven't changed. Until enterprises have full visibility and can reconfigure their networks easily, then the newest security trends will always come up short. Some are looking to APIs to solve this problem, but I believe the real breakthroughs will come through intelligent automation.

BN: With these trends, we're seeing the shift from NetOps to NetDevOps, can you explain what is NetDevOps and what it means for enterprises?

EL: NetDevOps is a collection of principles and strategies that encourage both people and operations to think more deeply about how network resources are consumed. It is defined by two key themes. The first is the introduction of DevOps principles that aim to break down siloes, which drives efficiencies by leveraging the best strengths across teams. The second is that NetDevOps works to provide a platform of capabilities that reduces manual work for teams.

In the short term, NetDevOps leads to more stable, secure and efficient networks, enabling IT and security teams to transfer basic, time-consuming tasks to automation so they can focus on improving business services, the customer experience, and higher-level processes.

BN: As network footprints expand, what are some of the issues that enterprises increasingly have to deal with?

EL: The challenge all networks deal with is complexity. Nobody set out to make the ecosystem of network standards, methodologies and hardware complicated, but a few periods of major technological transition have turned common tools into legacy infrastructure almost overnight. The result is that much of that legacy infrastructure is still embedded in our networks. Our current options are to make these problems go away through brute force overhauls or simply to automate around them intelligently.

BN: Widescale outages are certainly top of mind, what are some of the contributing factors that lead to the outages at Facebook, AWS, and Google Cloud in 2021?

EL: Enterprises like Google, Facebook and Amazon have an advantage in that they rarely have to deal with adapting legacy infrastructures to their needs because they have developed the hardware and networks for their specific use cases.

However, they face a different challenge of operating these custom networks at extreme scale. At this hyperscale level, companies are heavily reliant on automation to keep things running and the outages are often the result of misconfigured automation processes that cascade from something minor into an event that brings the whole network down. The fact that even these multinational enterprises struggle with automation despite their standardized infrastructures points to the challenge of achieving effective automation in-house.

BN: What are some of the pros and cons of self-built automation compared to off-the-shelf automation?

EL: Companies can take control of their own destiny with self-built automation because it can be built on its timeline, within its licensing requirements, and without locking into a specific vendor. Task-based automation becomes achievable, but many companies get so bogged down in deploying and keeping the automation up to date that they never actually benefit from the intended time savings.

So-called out-of-the-box offerings solve some of the big challenges easily, but the downsides come in at the back end. Off-the-shelf solutions require a company to buy in with a specific vendor, leaving them dependent on that vendor to implement changes or new features, which is often a slow and difficult process. Companies need a solution that addresses the big problems for them while also providing the ability to append and add features in a way that is customizable to maximize business value.

BN: How do you see the role of network automation changing alongside the evolution of IT operations?

EL: Today, companies pursuing increased digital transformation to improve their business services are getting caught up in how it will affect the network downstream. A self-operating, self-correcting network releases IT teams from focusing on the minutiae of hardware and software misconfigurations so they can instead focus on providing seamless business services. This level of automation requires a network utilizing AI, machine learning and RPA to contextualize anomalies, identify the relevant stakeholders, and suggest a fix so that IT teams can spend their time fixing the issue instead of trying to find it.

Self-operating networks implement efficiencies that enable IT teams to spend more time working on the big issues, such as collaborating with the business side to build upon their transformation journey to the benefit of the customer.

BN: Where do you think security falls into the mix?

EL: Security will remain a primary focus for all areas of IT, including the network. Ultimately, many day-to-day security concerns originate at the network level, whether it is network configurations, software currency, firewall rules, known vulnerabilities, and so on. As with IT operations, automation enables security teams to transcend the nuts and bolts of these basic security chores so they can focus on the integrity of overall processes.

BN: What are three things enterprises can do now to improve network operations and prepare for increasing scale?

EL: The first is to discover and maintain an accurate inventory of your entire network, establishing a single source of truth to operate effectively and minimize risk.

It is also critical that companies understand network configurations and have the ability to identify configuration drift across routing, switching, firewall, load-balancers, WAN optimizers and wireless LAN controllers, and vendor extensions.

Lastly, companies must maintain software currency to ensure that outdated versions don't leave networks open to unnecessary risk due to patchable vulnerabilities.

Image Credit: nmedia/Shutterstock

Comments are closed.

© 1998-2022 BetaNews, Inc. All Rights Reserved. Privacy Policy - Cookie Policy.