Many cloud-native companies have discovered that apps that require access to large data repositories, need exponential scalability or have been highly customised are more effective in a cloud-adjacent architecture. A cloud-only approach limits their ability to exchange data across their partner ecosystem without incurring unacceptable data egress fees. As a result, many of these cloud-native organisations are repatriating some workloads from public clouds to cloud-adjacent architectures.
Instead of connecting cloud workloads with traditional on-premises workloads via a WAN or the public internet—with all its security and performance issues—enterprises are placing their digital infrastructure in co-location centres. These facilities are also home to cloud on-ramps to leading IaaS and SaaS providers, B2B partners, and even customers. This allows organisations to bring their core applications directly adjacent to the cloud and their partner ecosystem, so that they can leverage software-defined interconnection and on-demand capacity for their next-generation digital services.
“More enterprises are standardising on cloud adjacency because it can help increase infrastructure agility, reduce latency, and keep sensitive personal data private and secure,” Jed Bleess, director of market development at Equinix, said. “In short, a cloud-adjacent architecture helps businesses create a hybrid multicloud environment that capitalises on the best of each element, truly delivering on the promise of hybrid.”
Service providers, including cloud providers, make up 64 per cent of the total mix for global interconnection bandwidth, compared with only 36 per cent for enterprises. The data also shows that service providers are digitally mature and expanding their infrastructure in a distributed manner: the average service provider deploys across eight core metros and eight edge metros. This demonstrates that they recognise the need to expand their reach to meet enterprise customers where they are.
“Enterprises are also expanding across core and edge locations to enable cloud adjacency,” Bleess continued. “In fact, the data shows Enterprises will grow interconnection bandwidth slightly faster than their service provider counterparts, 46 per cent CAGR by 2024, compared to 44 per cent for service providers. However, the GXI also states that over time, enterprises will grow their preference for as a service offerings leading infrastructure growth to be reflected more on the service provider side.
Among the key limitations of traditional on-premises infrastructure are the high costs and long lead times needed to stand up physical hardware. In the digital economy, where adapting quickly is one of the key criteria for success, businesses that can’t scale capacity on demand will get left in the dust.
Deploying adjacent to the cloud helps address this issue any time an unexpected traffic spike overwhelms reserved capacity within the data centre. Enterprises can use load balancing solutions from leading cloud providers to spread traffic across on-premises servers and virtual server instances that can be added or removed as demand fluctuates. Crucially, the virtual instances are connected to the on-premises infrastructure via secure, private interconnection, allowing them to function as a logical, albeit temporary, part of that infrastructure.
“The exciting possibilities of next-generation technologies like 5G wireless and the Internet of Things are another key reason cloud adjacency is more attractive than ever,” Bleess added. “These technologies are driving ever-lower latency caps, so it no longer makes sense for enterprises to operate traditional on-premises infrastructure that can’t connect with cloud services quickly and easily.”
Enterprises used to achieve ‘good enough’ latency results despite routing workloads through core hubs that were often hundreds of miles away. With use cases that require latency of less than 10 ms becoming more common, this is no longer tenable. Keeping workloads proximate to infrastructure is the only way to reliably keep latency low.
Use cases like smart manufacturing and connected vehicles use specialised hardware that typically wouldn’t be offered by cloud providers—such as sensors, cameras and robotic equipment—but can still benefit from the performance and flexibility benefits of cloud services. For this reason, these use cases are a particularly good fit for cloud adjacency: it allows enterprises to capture the data they need and put it into use quickly, without latency standing in the way.
“The digital world is increasingly global, populated by enterprises and service providers that operate worldwide,” Bleess said. “Governments are eager to prove they can keep citizens’ data private and secure in a world where national borders are increasingly blurry; as a result, countries across the world are stepping up data privacy regulations.
“Once again, cloud adjacency is the key: companies get the in-country physical infrastructure they need to satisfy regulators. They can also take advantage of cloud services without sending data out of country or exposing it to risk. Only data that needs to be moved to support a particular cloud workload will be moved, and any data that does pass to the cloud will do so over private, secure interconnection.”