Ciena thinks that cloud connect networks should be as elastic, flexible, scalable, and cost-effective as the cloud itself. But this requires a step-change in how these kinds of connections are built, operated and billed.

Michael Crossey is the Strategic Business Development Manager for Colt at Ciena.

According to Fortune Business Insights, the cloud market is expected to grow by CAGR 20% year-on-year, from US$677 billion in 2023, to more than US$2,400 billion in 2030.1

As cloud usage increases exponentially, businesses are also using an increasing number of different cloud providers. Indeed, figures from Statistic research suggest that 94% of large enterprises have already adopted a multi-cloud strategy, and that the majority (60%) of small businesses now use more than one cloud platform to power their businesses.

Accelerating cloud and multi-cloud adoption is unlocking major new commercial opportunities for businesses based on the ability to access new IT applications and services on demand, to convert large datasets into business insights, to connect distributed workforces with the software tools they need to work productively, and to provide more engaging experiences for end customers.

Along with benefitting businesses, the rapidly growing cloud market is also creating a wide range of new revenue opportunities for infrastructure and service providers of all kinds, from public cloud infrastructure providers to hyperscalers, content delivery networks, and network service providers.

To maximise these opportunities, high-speed network connections are needed that are capable of delivering large volumes of enterprise data to (and from) the cloud rapidly, cost-effectively, and reliably. However, this is much easier said than done, as historical approaches to cloud connectivity and cloud data migrations are typically slow, inflexible, and—in many cases—commercially untenable as well.

Why cloud-connectivity needs to evolve (and fast)

Traditional connectivity  services currently used for cloud access often have drawbacks when it comes to migrating large enterprise datasets from private on-site or collocated infrastructure to the cloud.

For example, many connectivity solutions typically require full payment for dedicated cloud port  usage and network bandwidth, whether or not the network is being fully utilised. Additionally, bandwidth available with this kind of service may not support larger connection rates, potentially slowing down an organisation’s cloud migrations and other projects—such as cloud-based analytics projects.

The second way to transfer large workloads and datasets to the cloud is via transfer appliances and services. But this approach is labour-intensive, costly, and slow, requiring data to be loaded onto physical drives and transported to data centres where cloud providers are located. Transfer appliances are also a one-way deal, limiting organisations’ options for moving between cloud providers or bringing data back in-house after analysis.

The third way businesses can move their workloads and data to the cloud is by using service provider custom connections. But this is also a restrictive approach . As well as requiring long (multi-year) contracts, this kind of connection provides fixed connectivity, with no support for bursting to support bigger-than-usual requirements.

For business customers wanting to migrate their data over a network to the cloud, traditional fixed networks can be fraught with time and cost risks. A company with 10  petabytes of data, for example, would need an always-on 10G  connection for 18 months to move its data one-way to the cloud. Transporting the same volume of data over a 100G connection would take the same business 56 days (around 1.5 months).

Bringing elasticity and scale to cloud connect networks

Fortunately for service providers, there is a new, better, faster, and more flexible way to connect businesses with the cloud platforms and providers of their choice. However, this requires a major step-change in thinking around how cloud connect networks are built, operated, and billed.

The current paradigm for service providers offering network connectivity services is to engineer their networks so that they are over-subscribed—meaning that the aggregate bandwidth offered to multiple customers is greater than the actual aggregate capacity of the network serving those customers.

This model has generally worked well in the past due to the ‘bursty’ nature of IP traffic, which means that customers very rarely need to consume their maximum available bandwidth at the same time. As a result, the network is generally able to cope with demand, without the full aggregate bandwidth limit being reached.

However, due to advances in technology that have improved both the cost and footprint per bit associated with networking equipment, a new paradigm has become possible. This involves over-provisioning the network to make large amounts of capacity available to customers that need to periodically move large datasets to the cloud rapidly and cost-effectively.

By overprovisioning edge/cloud connections in this way—typically up to 400 Gb/s but potentially to 800 Gb  /s or even 1.2 TB/s—service providers can introduce an entirely new network consumption model based on elastic usage of resources. This effectively brings the network in synch with the flexible, ‘pay-as-you-use’ model that has always been so attractive in the cloud computing model.

With this kind of elastic cloud-connect network, an enterprise would potentially pay a base monthly fee just for being connected. If they need to move a workload or dataset to (or from) the cloud at any time, they can simply request the bandwidth and duration of connection required from their service provider via a customer portal and pay the corresponding ‘usage-based’ fee.

The top benefits of elastic cloud connectivity

From an enterprise perspective, this new ‘elastic’ approach means that network services can be consumed like any other cloud resource, and provisioned in much the same way as compute and storage resources are provisioned in the cloud today. This, in turn, means that enterprises can gain a high level of self-service capability for accessing new network services, can access services faster, and can scale bandwidth up and down as their needs change—all with transparent, usage-based pricing.

Meanwhile, service providers can leverage their existing physical and fibre assets in the metro and edge networks to connect businesses with on-ramps between enterprise data centres and cloud edge nodes. This means that service providers can take on a much more strategic, value-added role at the intersection between hyperscalers and enterprises as demand for cloud services grows.

Enterprises will also be prepared to pay marginally higher rates for burstable bandwidth on demand as they will be able to reduce their expenditure on cloud port usage and pay only during periods of consumption. This is a win-win for enterprises—who may pay less overall for their cloud connectivity—and for service providers—who can increase revenues by supporting multiple customers on a single connection and maximising the commercial value of their network assets.

In terms of future opportunities, the availability of these kinds of cloud connections will remove the traditional bandwidth and cost barriers to enterprise migrations. This will result in more businesses sending more workloads and data to the cloud over time. This will especially be the case among smaller and medium businesses who are not yet ready to commit to fixed-bandwidth connections.

All of these benefits will drive overall market growth and create a wealth of new revenue opportunities for service providers, network providers, and all other organisations in the cloud value chain.

Make the move to elastic cloud networking with Ciena

Ciena has created a proven reference architecture that allows service providers to deploy burstable 400G to 800G cloud connections for multiple customers. We have a working architecture that woud underpin this move to elastic cloud networking and we are working with cloud providers to enable elastic cloud connections.  These architectures allow businesses to connect with the cloud platforms and services they need, and to migrate massive volumes of data to and from the cloud much more quickly and cost effectively than has previously been possible.

To discover more about the Ciena solution for elastic cloud connect networks, and how it can help you increase your cloud connectivity revenues and maximise lifetime returns on customer relationships, please read our new blog series (see below) or contact us today for a meeting and we’ll be happy to help.

The first blog in our series (Why service providers sit at the intersection of edge cloud drivers-and-demand) talks about many of the trends driving cloud services and how service providers are uniquely positioned at the intersection of enterprises and hyperscalers to capitalize on these trends. The second blog in our series (Its time for a new interconnection model for the growing amount of data in motion ) talks about the need for the network to support a consumption based model for ultra-high speed (100Gb/s or higher) cloud on-ramps to accelerate the migration of data to the cloud while aligning with the way that cloud services are consumed today.

1 Fortune Business Insights, “Market Research Report: Cloud Computing Market Size, Share, & COVID-19 Impact Analysis,” May 2023.