The recent outage at a prominent data center in Chicago has reignited the debate between colocation (Colo) and cloud-based architectures, especially regarding the disaster recovery mechanisms associated with these deployment models. Ironically, one of the coldest days of the year, the incident was caused by cooling issues. In response to this, the datacenter asked tenants to shut down their equipment which resulted in significant outages for numerous application and service providers.

Given the angst associated with the outage, it’s timely to review from a network perspective the Colo vs. cloud-based architecture debate with a particular focus on disaster recovery options available in both designs.

Before we deep dive into the approaches let’s first define the terms here.

Defining the Terms

What is Colo-Based Network Architecture?

In this model, enterprises house their routers, firewalls, servers, and storage hardware in rented space within a Colo data center, as opposed to maintaining these resources in- house or relying solely on cloud-based services.

What is Cloud-Based Network Architecture?

This refers to network infrastructures primarily utilizing cloud computing services and resources. In a cloud-based network architecture, enterprises leverage cloud-native network services or tools, along with network and security resources hosted and managed remotely on servers in data centers operated by cloud service providers.

Overview of Colo based architectures:

Colo based network architectures refer to network infrastructure which leverages Colo facilities to host the infrastructure. This allows enterprises to host their network and server infrastructure in a space managed by the Colo provider instead of maintaining and managing their own datacenter. This is a hybrid approach for enterprises where they don’t have to own a physical data center, power and cooling systems, but still own and manage the infrastructure devices like routers, firewalls, switches and servers. This architecture can offer scalability and flexibility, but it lacks agility and speed. For example, enterprises can scale their operations up or down by renting more or less space and using the Colo provider’s resources (like power, cooling, and bandwidth) as needed. However, they still have to manually order, install, connect, and configure the hardware. All of this prohibits agility. Alternatively, if you build your network for future scale and redundancy from day one, then you will be wasting all these resources without using them properly.

Colo provides an ideal solution for enterprises with unique application connectivity, scale, and security requirements that are difficult to meet with cloud-native constructs. It allows flexibility in implementing any design with any vendor of choice, without the hassles of maintaining data centers.

Disaster Recovery in Colo

Colo can offer a pretty robust disaster recovery plan as the Colo provider locations are dispersed globally. However, this usually involves maintaining at least two identical environments for failover, which can be costly. Businesses often face the dilemma of choosing between cost and the benefits of having a failover environment. In fact, not all enterprises can afford such redundancy. Oftentimes, they have to take the risk of running an environment without failover and trust the redundancy and robustness built into Colo provider’s data centers.

These Colo providers , no doubt have built a robust and reliable Colo based infrastructure. However, at the same time, they do occasionally experience outages similar to the one that happened in Chicago last week, causing disruptions for many enterprises who haven’t built failover locations for their applications nor had a proper disaster recovery plan in place.

Why Cloud Network Architectures Shine

In these circumstances, cloud network architecture shines because of its agility, low upfront cost, and the on-demand nature of the service. Some benefits include:

Scalability and Flexibility:

Cloud networks provide scalability and flexibility where resources can be scaled up or down almost instantly to meet changing demands.

Disaster Recovery

You can also extend this scalability and flexibility to other regions for disaster recovery. In the case of cloud, you don’t have to keep running your environment all the time. As part of your disaster recovery plan, you can create scripts using APIs or Terraform to quickly instantiate applications in another region in case there is an outage for a cloud region. This allows you to replicate the application without having to run it all the time, even when there is no outage.


The cloud offers a pay-as-you-go model, allowing businesses to pay only for the resources they use. This is better in contrast to significant upfront and ongoing costs associated with maintaining physical infrastructure in traditional data centers.

Global Reach

Cloud providers operate in data centers worldwide, offering businesses a global reach. This geographical diversity can significantly reduce latency and improve user experience, irrespective of the user’s location.

Focus on Core Business

With cloud architectures, businesses can focus more on their core operations and less on the complexities of managing IT infrastructure, which can lead to increased productivity and efficiency.


Cloud platforms provide the framework for deploying efficient and resilient architectures. However, this requires in-house expertise to define the right architecture based on business and application needs. This becomes even more challenging when dealing with multiple clouds, as each cloud provider is different and comes with its own limitations.

For enterprises facing these challenges, Alkira can help with our Cloud Networking Platform. We are a built-in cloud networking solution. We offer enterprises a platform which allows them to deploy a networking solution which is built for the cloud era. Through us you can onboard workloads hosted in public cloud providers and connect them with other cloud workloads, on-prem branches, data centers, and globally- distributed users without needing deep in-house cloud native networking expertise.

Enterprises leveraging Alkira will have most of their critical traffic traversing through our infrastructure and, as a result, the team here at Alkira takes extraordinary strides to make sure that our infrastructure is fully secure, highly available, and built to recover itself in different failure scenarios. Check out our blog, ‘Don’t Cut Corners on Cloud Network Resiliency’, explaining how we make our applications resilient to such failures and also provide you options to achieve the same for your applications.