Distributed Computing Is the Next Battleground, and Everyone Is Competing

Subscribe To Download This Insight

By Dimitris Mavrakis | 1Q 2021 | IN-6091

The B2B application domain has seen tremendous growth the past few years, with many different companies attacking it from different angles and with different strategic priorities. Hyperscalers, and particularly Amazon and Microsoft, are moving into the on-prem space with products like AWS Outposts, Snowcone and Snowball, and Azure with Private Edge Zones. Meanwhile, telecom operators are starting to target enterprises with hybrid connectivity models that include private cellular, network slicing and their crown jewel, Wide Area Network (WAN) connectivity. At the same time, there are startups entering the market with new products and services, that use components from both hyperscalers and telecom operators to implement new types of functionality.

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.

 

Containers, Micro-Services, 5G and Hybrid Clouds

NEWS


The B2B application domain has seen tremendous growth the past few years, with many different companies attacking it from different angles and with different strategic priorities. Hyperscalers, and particularly Amazon and Microsoft, are moving into the on-prem space with products like AWS Outposts, Snowcone and Snowball, and Azure with Private Edge Zones. Meanwhile, telecom operators are starting to target enterprises with hybrid connectivity models that include private cellular, network slicing and their crown jewel, Wide Area Network (WAN) connectivity. At the same time, there are startups entering the market with new products and services, that use components from both hyperscalers and telecom operators to implement new types of functionality.

In this rapidly expanding market, a new trend is emerging: distributed computing. This is an important development and will be the next phase of the hybrid cloud development, where processing payloads will be distributed between public, telco, edge and on-prem clouds.  

The Rise of Distributed Computing

IMPACT


In distributed computing, not only processing payloads are distributed, but also mobility and handover of these payloads between different cloud domains are possible, and in many use cases, necessary. For example, an enterprise that deploys a containerized B2B application in its on-prem cloud may require short-term processing and storage requirements it cannot support. As such, this application may utilize the public cloud for heavy processing, the telco cloud for specific applications and the edge cloud for low-latency or near-real-time applications. Another example could be autonomous driving or Autonomous Ground Vehicles (AGVs) that rely on edge resources for certain time sensitive and compute intensive workloads. In this scenario, when these vehicles are mobile, the payload should be handed over to neighboring edge computing hubs as they navigate their journey.  

Although hyperscalers and mobile operators have already launched products and services that tie two of these domains together, most of them are static and only designed to take advantage of the specific domains they were designed for. For example, an enterprise that utilizes AWS wavelength deployed on a telco cloud for its applications is not able to migrate payloads between another public cloud or their own private cloud, let alone in appliances that are outside the sphere of the AWS ecosystem.

This is the challenge that distributed computing can enable, the democratization of processing payloads and free migration of applications between different domains. In a way, distributed computing will break free from the hyperscaler monopoly and allow enterprises to implement efficient cloud models that are designed for their specific use cases. However, there are significant challenges before this becomes a reality. For example, there is no open (i.e. not controlled by a single player) orchestration framework that can manage these payloads and sync resources between different computing domains, but market activities are moving towards this direction.

Market Activities for Distributed Computing

RECOMMENDATIONS


The market is now exploring advanced cloud architectures and business models , exemplified by the emergence of the Open Distributed Infrastructure Management (ODIM) project, that aims to specify an open source software platform that delivers means of distributed physical infrastructure lifecycle management, across telecom and enterprise edge-to-core  infrastructure deployments. Moreover, there are several startups that utilize hyperscaler and edge platforms to allow developers to distribute their payloads in a more geographically diverse deployment compared to previous deployments. One of these companies is Macrometa, giving developers the flexibility to build real-time applications and APIs on its secure, programmable stateful-serverless platform. Intel has also launched the OpenNESS platform that aims to orchestrate processing payloads between enterprise and edge clouds for enterprise applications.

Naturally, hyperscalers are already present in this market and in a way, dominating. An enterprise that uses the AWS cloud and either an edge (e.g. Wavelength) or on-prem (Snowcone or Snowball) is already utilizing distributed computing, but locked into the AWS ecosystem. If this enterprise aims to deploy an application that is not running in the AWS ecosystem, they may not have a choice to deploy it. This is what the market is currently aiming to solve, and putting telecom operators in a good position, since they are already operating distributed computing platforms, their networks, and are familiar with managing and orchestrating them.

The distributed computing opporunity is vast, and will likely be the next wave of competition and innovation. The spheres of influence are not yet set in stone, and the next few years will be pivotal for many established and new companies.

If you would like to learn about the market development around Distributed computing, please refer to our newly launched Distributed and Edge Computing Service.

 

Services