In 1994 L. Peter Deutsch presented his ‘Fallacies of Distributed Computing‘ and spoke about the false assumptions made by developers when writing distributed applications. Understanding these fallacies is key, because in 2022 software development in distributed IT is more important than ever before. Network experts have to ensure that networks support a broader spread while also ensuring that networks serve as a secure foundation for other trends, such as edge computing and data fabric architectures. The key to success is to understand the importance of virtualized networks and the ability to monitor networks in depth.

It is unlikely that companies will return to an office-centric approach to work after the massive proliferation of remote and hybrid working in 2021. According to the industry analyst Gartner, the main challenge for companies in 2022 will be to become ‘distributed enterprises’. A speaker from Gartner argued that every organization, from retail to education, will need to reconfigure its delivery model to embrace distributed services. Gartner expects that by 2023, 75% of organizations that exploit the benefits of distributed enterprises will realize revenue growth 25% faster than competitors.

Assuming that companies will become more geographically-distributed, you can expect that, on the one hand, companies will make further use of cloud technologies. These are easier to scale and have other benefits. But, besides the well-known advantages of the cloud, organizations will, on the other hand, decentralize and accelerate the adoption of edge technologies. Research by IDC expects a 40% increase in edge network deployment by 2022. IDC also reveals that the top three reasons for putting more effort into the network edge are to improve productivity and efficiency, security and compliance, and to accelerate decision-making.

While it sounds confusing at first that companies centralize some assets in the cloud, and at the same time invest in decentralized IT on the edge, this does actually make sense, if you look at the broader picture. Increased cloud usage and edge computing tied together by a new type of application architecture are extremely powerful when it comes to improving and delivering better services. Data fabric approaches have the potential to become such a unified architecture to cover hybrid infrastructures, but it will probably take a few more years until data fabrics reach a ‘Plateau of Productivity’.

Virtualized networks and beyond

Regardless of how quickly organizations aim to adopt all these trends, they all put a burden on networks: Corporate networks will be stretched out, have to carry more data and will be more prone to security incidents. Investments in reliable networking hardware with a robust physical architecture are no longer sufficient to deal with these challenges. Thus, in 2017 Gartner had already announced the shift from ‘software-defined networks’ to ‘intent-based networking’ as 'the next big thing'.

The challenge is that most companies use different technologies to virtualize their networks. Intent-based networking aims for automated, intelligent networks. Server virtualization and segmentation in VLANs create software-defined networks, but these are both fairly far from such an approach. It is difficult to transform networks, especially if their reliability, performance and security are falsely taken for granted. A transformation towards automated, intelligent networks requires a significant network-topology readjustment. Requirements such as switching from a port-centric design to a policy-centric network, the creation of microsegmentation and feedback mechanisms for the adoption of intent-based networking will take time. Because there is no ‘out of the box’ solution, an iterative process is recommended, and network architects will need to take several steps towards the right direction in 2022.

A first step for dealing with the further distribution of networks is the building of secure tunnels and managing these efficiently. Tunnels are important to share resources when working remotely. The security of VPNs can be improved by approaches such as Internet Protocol Security (IPsec). Automation tools like Ansible are a way to automate the configuration of network devices like firewalls or switches, including their routing policies, so that only necessary traffic is routed through company VPNs. The combination of IPsec and automated configuration management is just a suggestion to improve security and reduce the workload of network architects at the same time. But clearly, companies have to find ways to optimize their virtual tunneling.

Organizations will also probably shift away from Multiprotocol Label Switching (MPLS) to approaches like software-defined wide-area networks (SD-WANs), if possible. MPLS does not include encryption and is also becoming problematic, because all traffic from all local networks usually is routed through static and laborious routes. Often data packets are always sent to a central node, which pushes a lot of traffic on the carrier. SD-WANs allow a more flexible routing of network traffic between several locations. Because a SD-WAN relies on intelligent firewall policies to route traffic depending on the device and application from a system directly to the target, network architects have to spend more time on the configuration of network devices. To deal with the additional workload, you have another argument for why the automation of configuration in network devices will gain more attention in 2022.

Regarding security, Gartner expects that ‘Zero Trust Architectures’ (ZTA) in networking will become more common. Instead of just user credentials such as a password, a proxy decides whether to grant each access request based on a risk profile calculated from combined contextual factors. In May 2021, the US government published an executive order asking federal agencies to adhere to Zero Trust. This security paradigm is not entirely new, but as environments are becoming highly diverse and distributed, IT teams outside the US government sector should also prepare for implementing ZTA in 2022.

Network monitoring in 2022

Monitoring will not stop the growing volumes of data and the further spread of networks. But, monitoring all devices in an IT infrastructure is the best way to guarantee uptime and performance. Precise monitoring data is also the basis for prioritizing modernization efforts in networks, especially if your monitoring includes forecasting.

Some network monitoring approaches are now already unable to cover all components and relevant parameters such as packet rates, error rates, bandwidth, and the state of the ports on the various switches and routers. Thus, network architects need to make sure that they have a state-of-the art approach that goes beyond SNMP, but which also allows in-depth analysis of their network. For example, monitoring network flows allows root cause analysis, helping to identify bottlenecks in the network more quickly, an ability more important than ever in 2022.

Automation in monitoring is also crucial. It does not help if a network architect is automating the management of network configurations, but has to configure the monitoring manually.
Switching to SD-WANs, for example, will require the implementation of various modern firewall appliances. A monitoring needs to support such high-end network devices, ideally leveraging vendor-specific APIs, and also has to add these systems automatically into the monitoring and collect data more efficiently. The monitoring also has to be able to evaluate new network approaches and detect any misconfiguration immediately.

The further distribution and the focus on edge environments also demand matching monitoring approaches that allow short ways for the monitoring data and shift monitoring workloads as close to the monitored assets as possible. Edge devices keep more workloads in the local networks and make essential processes more resilient. These benefits are reduced with cloud-based monitoring approaches that need extra bandwidth. Distributed monitoring environments are ideal for monitoring edge devices, because they are easy to set up and manage, but also reduce the network bandwidth used for monitoring. Also, if the internet connection goes down, the local monitoring still continues to run.

Because remote work will continue to be on the agenda, a monitoring solution needs to provide full flexibility to monitor the various technologies essential for remote access. VPN monitoring is a must, but new technologies like ZTA have an impact on monitoring, too. Because not all assets will be able to support ZTA, for example, and the National Institute of Standards and Technology (NIST) recommends creating segments for such applications. These segments should only be accessed through secure gateways. An IT monitoring needs to be able to monitor such segments and their gateways, of course.

Take-aways

Having a misguided faith in networks has led to disasters several times. In 2022, not paying attention to the fallacies of distributed monitoring will become an even bigger deal, because most IT innovations assume that networks run at peak performance even when they are stretched out like never before.

Network architects have to manage several mission-critical tasks. First of all, they need to make sure that corporate networks are running and deliver the resources expected. Because assets are more distributed and the general volume of data is growing, this can be challenging. Most network architectures cannot yet provide a fully-automated configuration of resources. A monitoring tool needs to provide precise data to be able find bottlenecks and ensure compliance with industry standards.

Secondly, architects will need to make changes in their networks and to monitor new technologies. For example, edge devices are getting more and more common and also need to be part of a network monitoring. A monitoring tool needs to be able to support such modern devices with as little effort as possible.

Thirdly, network architects have to understand the mission-critical roles of networks in 2022 and beyond. Organizations can only implement important future trends such as data fabric approaches when their hybrid network architecture is suited to it. Given the complexity of network operations, it is important to take the right steps as soon as possible.