Blog

Edge Primer: Distributed Cloud – The power of public cloud at the edge

By July 27, 2021No Comments
By Utpal Mangla VP & Senior Partner; Global Leader: IBM’s Telecom Media Entertainment Industry Center of Competence; 
Ashek Mahmood, IoT Edge AI – Partner & Practice Lead, IBM; and Luca Marchi, Associate Partner – Innovation Leader – IBM’s Telecom Media Entertainment Industry Center of Competence

Today, many enterprises benefit from using public cloud to build and run their applications.  Consuming IT as a service, through APIs, at scale, and as needed, has transformed how applications are built. Public Cloud enables companies to innovate in new, faster ways. But the reality is only somewhere between 5% to 20% of enterprise workloads have moved to the cloud. 

That first wave of applications was focused mostly on new workloads, or workloads that can easily relocate. Many applications have requirements on security, compliance, latency, regulation, and performance and cannot easily move into a public cloud. Enterprises need a way to gain the benefits of public cloud anywhere they are running their applications.

The solution that helps is called “Distributed Cloud.” Distributed Cloud is a new cloud computing model that extends public cloud services to any location, even at the edge. This means companies can now deploy cloud-native applications not only in their primary cloud provider’s infrastructure but on premises, within other cloud providers’, in third-party data centers or colocation centers, or at the edge – in factories, distribution centers, stores, hospitals and ports – with everything managed from a single control plane. 

Distributed Cloud provides greater control to an organization of its hybrid cloud ecosystem. It mitigates dependency on one provider and provides flexibility to organizations to choose, based on specific business requirements. With Distributed Cloud, developers can take advantage of the catalog of services from public cloud providers with industry-optimized security and compliance and leading AI capabilities anywhere its needed. All with a common API, user experience, management dashboard and user consumption model. 

With 5G use cases across industries becoming a reality, massive growth of IoT devices (in trillions of volumes) is expected. Control and audit of network traffic is critical for maintaining data protection, privacy, and compliance. Management from a single control enables zero-trust Edge security across all devices and users, irrespective of which network they are on. Distributed cloud extends the capabilities through multiple cloud satellites across edge networks. 

To get a better idea of how Distributed Cloud works, let’s look at some actual applications from different industries.

  • The first application comes from the financial services industry. In order to compete in a market being disrupted by newcomers, financial institutions need to create exceptional user experiences powered by technology. To do that, they leverage AI and trading algorithms to stay ahead of the market. The competitiveness of financial services organizations is based on quick iteration and deployment of cloud-native tools and best practices. However, the financial industry is heavily regulated. So, they need to be able to keep that data secure and compliant – which often means there are restrictions on where that data can live. 
    • Distributed Cloud can help solve this conflict. These financial services companies can now extend public cloud services in their data centers, allowing them to leverage cloud native best practices and meet their security and compliance obligations, all at the same time. So, Distributed Cloud enables financial institutions to take advantage of a public cloud consumption model, but in many different locations – and they do not need a high level of skills to run this software outside of the public cloud. 
  • Another example is the construction industry managing its worker health and safety regulations. One key player in the industry is leveraging distributed analytics to keep workers safe. In this scenario, a builder might have a business, or an office building that is under construction, and they need to leverage video analytics to tell if someone is wearing a hard hat or not – and warn them before they move into a potentially dangerous area of that building. They want to use video analytics to automatically detect and alert this situation, but latency could be a real problem – delays in such alert could fail to prevent the injury. 
    • To solve this safety problem, the construction company needs to analyze video feed from many cameras all over the office building, but it is impractical to send all of that data back to the cloud to be processed remotely. It would be better if the video processing happened close to the actual device. Traditionally, that would require you to install servers and software and manage them on-prem.   
    • With Distribute Cloud, the AI Video processing is extended into the office, so cloud services can be leveraged to run this application close to the device – and that is critical in ensuring that latency is reduced, and workers are effectively warned before they enter a dangerous area. 
    • That same company had to adjust to challenges brought by the Covid pandemic. They were able to quickly iterate on their application and modify the use-case for AI models and rapid deployment to distributed locations. Now, instead of warning workers about wearing hard hats, they can make sure that workers are wearing masks and ensure that the masks are being worn correctly. Additionally, they can even leverage thermal devices to take temperatures. 

Distributed Cloud enables organizations to quickly address unforeseen challenges, leverage cloud benefits in any location, and innovate quickly. 

See how the Akraino project’s StarlingX Far Edge Distributed Cloud blueprint can serve as a starting point: https://wiki.akraino.org/display/AK/StarlingX+Far+Edge+Distributed+Cloud

####