What is a distributed cloud?
A distributed cloud is an architecture where multiple clouds are used to meet compliance needs, performance requirements, or support edge computing while being centrally managed from the public cloud provider.
In essence, a distributed cloud service is a public cloud that runs in multiple locations, including
- The public cloud provider’s infrastructure
- On-premises at end customer locations in the data center or at the edge
- In another cloud provider’s data center
- On third party or colocation center hardware
Although there are multiple locations and geographies involved, all of the cloud services are managed as on from a single control plane that handles the differences and inconsistencies in such a hybrid, multi-cloud environment.
This distribution of services enables an organization to meet very specific requirements for response time and performance, regulatory or governance compliance mandate, or other demand requiring cloud infrastructure to be located anywhere other than the cloud provider’s typical availability zones.
The growth of the internet of things (IoT) and edge computing have been a major driver for distributed cloud deployments. Artificial intelligence (AI) applications that move large amounts of data from edge locations to the cloud require cloud services to be close as possible to edge locations, and moving cloud resources to the edge location itself can greatly increase performance for these applications.
Additionally, the ever-increasing number of government regulations such as the EU’s GDPR can demand that data be located in specific jurisdictions which may or may not be supported by a given public cloud provider, thus making a distributed cloud a necessity.
By bringing cloud services closer to a given user, application, or data, distributed clouds can offer
- Reduced latency
- Lowered or eliminated network congestion
- Guaranteed quality of service (QoS) for mission critical application and mobile users
How does a distributed cloud work?
In a distributed cloud, services are located or ‘distributed’ to specific locations to reduce latency and these services enjoy a single, consistent control place across public and private cloud environments. Gartner states that organizations can see major performance gains by lowering latency and reducing the can deliver major improvements in performance due to the elimination of latency issues, reducing overall risk of outage or control plane inefficiencies.
A distributed cloud takes not just an application but the entire computing stack and distributes it to the locations where it is needed, whether public cloud provider, on-premises, or in third party colocation facility. The consuming cloud customer sees this distributed infrastructure as a single cloud entity, and the cloud provider manages all of the elements of the distributed cloud as a whole from a single control plane.
The public cloud provider continues to be responsible for all cloud operations, including security, availability, updates, and governance of the entire distributed infrastructure. To paraphrase Gartner, distributed cloud fixes what hybrid cloud and multi-cloud breaks.
What are use cases for distributed clouds?
Distributed clouds offer a broad range of applications from smart edge computing to simplifying management of multi-cloud environments and hybrid deployments. Common use cases include:
Distributed cloud and edge computing support everything from simplified multicloud management, to improved scalability and development velocity, to deployment of state-of-the-art automation and decision support applications and functionality.
- Edge/IoT. With new uses for video inference and facial recognition being developed daily, IoT is using AI and machine learning (ML) for improving automobile manufacture, analyzing medical imaging, to smart buildings and smart cities that find the shortest route to parking and turn off the heating after the last employee has left for the evening. Many of these applications would be stymied if data had to traverse from the edge back to the cloud or a data center for analysis and processing.
- Content optimization. Distributed clouds can effectively become a content delivery network (CDN), which can improve the streaming experience or reduce web page load time latency, offering the best possible user experience for a broad range of applications.
- Scaling on demand. Distributed clouds enables expansion to existing locations without the need to build out additional infrastructure. As needs grow, the cloud footprint can grow along seamlessly, supporting the organization’s changing needs.
- Single pane of glass management. Adopting a distributed cloud approach gives enhances visibility into a hybrid, multi-cloud deployment, including the ability to manage all the infrastructure as a single cloud from a single console with a single set of tools.
- Meet compliance mandates. Local, federal, and international data privacy regulations can mandate where a user’s personal information must be stored, and whether that information can travel outside that jurisdiction. In those cases where the data can’t be moved to the public cloud provider, the public cloud provider can effectively be moved to the data ensuring not only that governance and regulatory mandates are met, but that data will be processed as efficiently and with the minimum amount of latency.
What are the benefits of distributed cloud?
There are many benefits of a distribute cloud architecture. Gartner points to these as noteworthy:
- Increased compliance. Distributed by nature, workloads and data can be located where they must be to meet regulatory demands.
- Increased uptime. Since cloud services can reside on local subnets they can be isolated – even untethered from the main cloud – when needed to ensure they are isolated from a crashed system to provide redundancy.
- Scalability: Adding VMs or nodes as needed not only enables rapid scalability, it also improves the overall availability of the cloud system as a whole.
- Flexibility: Distributed clouds simplify installation, deployment, and debugging of new services.
- Faster processing. Distributed systems can be faster by leveraging compute of multiple systems for a given task. Also, the distributed cloud enables more responsive communications for specific regions.
- Performance. Unlike centralized computer network clusters, the distributed cloud can provide higher performance and better cost performance.
What is the difference between cloud and distributed cloud?
- Traditional Cloud Computing is the delivery of IT resources and services on-demand, including servers, storage, and databases to name a few. These services are typically provided over the public internet or private network connection from one of many hyperscale cloud providers. Cloud services can be categorized as public cloud, private cloud (including on-premises data centers), hybrid cloud (the combination of public and private) and multi-cloud (including multiple public cloud providers).
- Distributed Cloud Computing discards the categories of public, private, hybrid, and multi-cloud. The distributed cloud presents to the user organization as a single cloud platform, but in reality it is comprised of multiple components that can include ‘all of the above’ – public cloud elements from the primary provider and one or more of its competitors, private cloud or enterprise data center, and third party colocation partner. These varied elements are all managed as one by the primary cloud provider and consumed as one by the ultimate customer.
What are the challenges of distributed cloud?
Managing an enterprise using a multi-site cloud deployment has its challenges including:
- Bandwidth. A broadly dispersed multi-cloud environment may have many different connectivity models for each location. As such, moving more computing to the edge can stress existing broadband connections and require upgrading or adapting to meet increased demand for throughput.
- Security. Securing a distributed cloud presents new challenges for both cloud provider and end user, as resources can be scattered across the globe and can be collocated with other enterprise server and storage resources
- Data protection. Backup and business continuity plans for dispersed data resource may require a redesign of backup and recovery strategies to ensure data stays in the geographies it is supposed to.