By Saju Sankarankutty
While cloud computing has fuelled digital transformation in enterprises over the years, the focus in the early stages was primarily on addressing migration, integration, and commercial challenges, and mitigating cloud adoption risks. As they progress in their cloud journeys, it’s time to move on from shopping and subscribing to exotic services from the cloud providers to building an end-to-end business solution that delivers value to the users.
Cloud network is one such area that deserves our attention. It connects various cloud services hosted across multicloud. Typically, enterprises focus on the last mile connectivity between the enterprise and the cloud network and largely overlook intra-cloud networks. They assume that the hyperscalers will take care of the network aspects of the service offerings and provide unlimited capacity. However, as enterprises increasingly deploy complex digital applications across clouds with stringent performance requirements, cloud network performance becomes a key factor in meeting response time expectations.
A robust multicloud network management enhances the value of cloud
Multicloud network management has become an important aspect of the overall cloud management framework since a considerable amount of business workloads are getting migrated to the cloud. Focusing on multicloud network management can help enterprises accrue several benefits including:
Consistent user experience
With hybrid work becoming a norm, the average user experience journey traverses across their home/office network, WAN, and cloud network. All these three network segments must perform within the expected capacity. Performance degradation in any of the networks can impact user experience.
Security and observability
From an observability and manageability perspective, it’s important that the network support team has an overall view of the cloud and enterprise networks under one unified pane. This can not only help identify the bottlenecks but also improve support metrics such as Mean Time to Detect (MTTD) and Mean Time to Respond (MTTR).
From a cyber security perspective, network devices such as routers and access points are often the first attack points. These devices need to be patched in a timely manner to reduce the blast radius. With critical applications moving to the cloud, similar patching needs to be done on the cloud network resources such as transit gateway, internet gateway, and VPN gateway. The patch solutions for cloud and non-cloud network components often vary. Hence patching network devices across cloud and on-premises becomes time-consuming and prone to errors or gaps, leading to open vulnerabilities. A unified patching solution to patch the entire network devices is a must for organisations that are heavily invested in the cloud.
Better network management
While automation is used significantly in driving operational efficiency, its use is severely limited in enterprise LAN and WAN services. The advancement in software-defined techniques such as SDN, and SDWAN have brought agility in configuring and managing networks remotely. Most of the automation tools are provided by the OEM and they automate only the devices belonging to the respective OEMs. Unified cross-OEM automation solutions are largely missing in the commercially available automation toolsets. Enterprises, therefore, need to put in place a unified platform to automate the network as a whole.
Unlike traditional infrastructure where resources such as servers, VM, and storage are dedicated to a particular business unit, network resources are typically shared across several business units. In a scenario where IT wants to offer a network as a service to a business, it’s important to amortize the network cost based on certain business logic such as consumption, number of users, number of VIP users, applications accessed, and more. Having a single unified platform that can provide comprehensive insights and reports into network cost consumption of cloud and non-cloud networks, based on location, business unit, device type, and other factors would help organisations to perform cost allocation and budgeting with efficiency.
Capacity management of network devices is yet another critical aspect that needs to be considered while building and managing a complex network landscape since any performance issue in a network ingress/egress point can impact a large user population. Applying AI/ML technologies on telemetry data collected from various network devices can help organisations gain detailed insights and take necessary actions. For instance, organisations after studying the underlying network device utilisation can proactively initiate change management process flow to upgrade the device and alleviate potential performance issue
It is crucial that all network cloud infrastructure is compliant with the organisation’s standards and adheres to regulatory requirements such as GDPR, SOX, HIPPA, and others. Having a unified portal that can detect the compliance drift of cloud and non-cloud network infrastructure effectively and fix them in an automated fashion is a must for all organisations.
As the cloud becomes all-pervasive and organisations become increasingly reliant on the cloud for driving innovation and delivering a solution to the business, it’s important to have capabilities to orchestrate and manage the underlying network infrastructure efficiently.
(The author is Vice President and Unit Technology Officer, Cloud, Infosys. Views are personal.)