Internet service providers (ISPs) are the backbone of modern society, responsible for connecting businesses, services, and people to the Internet and to each other. ISP networks are vast, distributed, and complex, making them challenging to manage effectively. However, failing to do so has major consequences. For example, in July of 2022,
Rogers Communications in Canada suffered a network system failure after a maintenance update, causing an outage that lasted more than 15 hours and took down emergency services and other critical infrastructure.
An ISP network architecture must be designed for resilience to prevent major incidents from occurring that affect consumers, communities, and the provider’s reputation. But significant challenges stand in the way, including a reliance on legacy infrastructure, and an inability to troubleshoot and recover failed gear remotely. This post discusses why these challenges exist and what ISPs can do to overcome them.
ISP network architecture challenges
Many ISP networks lack resilience because providers are failing to adapt to a rapidly changing landscape. With networks growing larger and more complex every day, new technologies like AI (artificial intelligence) and software-defined networking are needed to manage infrastructure efficiently and deliver innovative services. Additionally, providers get stuck in a break-fix cycle that leaves teams struggling to maintain service level agreements or focus on innovation. Let’s look at the causes of these challenges and discuss how to build more resilient ISP network architectures.
Legacy infrastructure creates technical debt and hampers growth
|
The challenge:
|
The solution:
|
|
Reliance on legacy systems creates technical debt and prevents ISPs from implementing new technologies
|
Vendor-neutral platforms like Gen 3 serial consoles extend automation, software-defined networking, and other advanced technologies to legacy infrastructure until it can be replaced.
|
Internet service providers often have a network architecture that’s a mix of new and legacy infrastructure. However, engineers with the experience to support older solutions are no longer working in the field, either because they’ve been promoted to leadership positions or retired. When legacy hardware fails, inexperienced engineers need time to overcome this skills gap, and ISPs may even need to bring in consultants. This increases the cost of failures, creating what’s known as “technical debt” – when a solution is more expensive to support than the value it brings to the organization.
In addition, ISPs can improve network resilience and provide better service to customers, by adopting new technologies like AI, 5G, software-defined networking (SDN), and Network as a Service (NaaS). But legacy hardware hampers the ability to adopt these technologies. For example, NaaS abstracts the need for MPLS circuits and customer-premises gear, making architectures more cost-effective and improving the customer experience. NaaS brings SDN concepts like programmable networking and API-based operations to WAN & LAN services, hybrid cloud, Private Network Interconnect, and internet exchange points. It optimizes resource allocation by considering network and computing resources as a unified whole and attempts to automate as much as possible. The trouble is, ISPs struggle to implement NaaS and other beneficial new technologies because their legacy hardware simply can’t support it.
Solution: Legacy modernization with a vendor-neutral platform
The ideal solution is to replace legacy infrastructure with modern hardware and software that supports the latest technologies. But for many ISPs, an overhaul like this is too costly and intensive. The next-best option is to bridge the gap with a vendor-neutral network modernization platform that extends automation, AI, and 5G connectivity to otherwise unsupported systems.
For example, serial consoles (also known as terminal servers, console servers, and serial console switches) provide remote management access to network infrastructure. The newest generation of these devices, known as Gen 3, are vendor-neutral by design so that they can control third-party and legacy hardware. Through a combination of built-in features and integrations, Gen 3 serial consoles can use technology like zero-touch provisioning (ZTP), AIOps, and automated configuration management to control connected hardware that otherwise wouldn’t support it. Some solutions, such as the Nodegrid platform from ZPE Systems, can even directly host SDN and NaaS software from other vendors, so ISPs can start implementing network improvements right away while they gradually replace their outdated infrastructure.
Physical infrastructure is difficult to manage and troubleshoot remotely
|
The challenge:
|
The solution:
|
|
ISP network admins can’t respond to changing environmental conditions or recover failed hardware remotely
|
Environmental monitoring connected to an out-of-band (OOB) management solution ensures continuous remote access on a dedicated, isolated network that enables fast and cost-effective recovery.
|
ISP network architectures involve a great deal of physical infrastructure, which is often deployed in remote edge sites and customer premises. Even with software- or service-based network solutions, hardware is needed to host that software, and the physical environment for that hardware is often less than ideal. Drastic weather changes, power outages, and other unexpected scenarios can happen without notice and rapidly bring down an ISP network. These events often cut off remote management access as well, making troubleshooting and recovery difficult, time-consuming, and expensive. In fact, supporting this physical infrastructure often consumes so much time and effort that it prevents ISPs from focusing on delivering better services and software to their customers.
Solution: Out-of-band management with environmental monitoring
The first part of the solution involves monitoring the environment that houses remote, physical infrastructure. An environmental monitoring system uses sensors to detect changes in airflow, temperature, humidity, and other conditions that affect the operation of network hardware. These sensors give ISPs a virtual presence in edge deployments and customer sites so they can quickly respond to changing conditions before systems overheat or circuitry corrodes.
The second part involves providing management teams with reliable remote access to physical infrastructure that won’t go down if there’s a production network outage. Out-of-band (OOB) management solutions use serial consoles with dedicated network interfaces used just for management access. This creates a parallel, out-of-band network that’s completely isolated from production network services and infrastructure. Additionally, many serial consoles use cellular connectivity via 4G or 5G to OOB access, providing a wireless lifeline to connect, troubleshoot, and restore remote infrastructure. OOB management allows ISPs to troubleshoot and recover failed hardware remotely, even during total network outages, so they can get services back up and running faster and less expensively.
The environmental monitoring system should run on the OOB network so remote admins can continue to monitor conditions while they recover failed hardware. The out-of-band management solution also needs to be vendor-neutral so ISPs can deploy third-party automation, AI, and NaaS on the OOB network. For example, Nodegrid Gen 3 serial consoles provide OOB, environmental monitoring, and a vendor-neutral platform to host third-party software at the edge. Nodegrid even enables fully automated responses to changing environmental conditions in those edge environments before admins are aware of a problem.
