The rise of Kubernetes for enterprise edge computing
By Brooke Frischemeier Head of Product Management, Unified Cloud Rakuten Symphony
June 4, 2024
Over the history of data centers, we have seen trends come and go from on-premise, to client-server, to distributed systems and every possible term thrown in front of the word cloud, each of which is marketed as “the next big thing.” We have seen these changes and upgrades driven by numerous technological advances that created better business models, driven by facilities cost, Wide Area Networking (WAN) price-performance, mobile networking, chip-sets economy of scale, staff out-sourcing, data center repatriation and the list goes on and on. Fortunately, these models, along with their own lessons and gains are incremental, and it has led us to modern-day edge computing for many key verticals including retail, finance, AI/ML, Industry 4.0 and content delivery, just to name a few.
How enterprise edge changes the paradigm
Edge computing seeks to revolutionizing how data is analyzed, processed, stored, and acted upon. It represents a significant departure from the conventional cloud-centric approach by shifting compute resources closer to the data’s source, all the way to the "edge" of the network, closer to the user. This fundamental change to processing locality is ushering in a new era of responsiveness, efficiency, and scalability, with Kubernetes playing a key role in delivering this transformation.
At its core, edge computing is about decentralizing and distributing computing power and data processing, moving it from centralized cloud data centers, closer to the source of the action. In edge computing, data is processed and analyzed locally, without the need to send it over costly, slow WAN links, to central, far away data centers. This proximity to data sources significantly reduces latency and data bloat, ensuring real-time or near-real-time responses while reducing capital and operational expenses. It’s a win-win.
What are the key characteristics of edge computing?
Real-time processing: As we alluded to earlier, real-time processing is the hallmark of edge computing. It enables applications to make instant decisions based on data from sensors, cameras, and other sources, supporting smart cities, logistics, manufacturing and Internet of Things (IoT).
Reduce latency: Another challenge of the past has been latency, the bane of numerous applications over the years, and it is one of the most critical areas edge computing is helping to reduce. With edge computing we can drastically reduce round-trip response time and do so predictably. Many applications, such as augmented reality/virtual reality (AR/VR), autonomous vehicles, and remote experts require minimal latency to provide safe, seamless and immersive user experiences.
Data localization: With edge computing, data stays closer to its source, reducing the need for extensive data transfers and optimizing bandwidth usage. Data localization is particularly advantageous for remote or bandwidth-constrained locations. This in turn builds on lower latency/response times, by capitalizing local actions in industry, enterprises, content delivery and interactive networks.
Decentralized compute: Since edge computing decentralizes computational resources, distributing them across the network, it adds efficiency, fault tolerance and availability to the user experience. Centralized cloud services are vulnerable to network outages and latency fluctuations. Edge computing enhances reliability by allowing applications to continue functioning even when connectivity to the cloud is disrupted.
Scalability: Not all centralized clouds can be easily expanded upon, especially when you own and operate it yourself. Edge computing architectures are inherently scalable, as additional edge nodes and resources can be easily added, accommodating increased workloads and data volume.
Bandwidth optimization: Lastly, edge computing enables bandwidth optimization. Since data can be processed and reduced by edge applications, it reduces the strain on network bandwidth. Furthermore, it reduces the total amount of data that needs to be stored in a central data warehouse, reducing the costs of networking and storage resources.
In conclusion, edge computing is a transformative computing paradigm that brings data processing and decision-making closer to where it's needed most. With Kubernetes facilitating the orchestration and management of edge workloads, it is poised to play a pivotal role in enabling real-time, scalable, and responsive applications across a wide range of industries and use cases.
Kubernetes for the win!
Kubernetes has emerged as a compelling solution for edge computing due to several key factors that make it well-suited for managing distributed resources. Kubernetes comes with key set of advantages that align with the unique requirements and challenges posed by edge computing.
Kubernetes is an orchestration system designed to manage containerized workloads. Containers are lightweight, portable, and conducive to microservices architecture. This enables one to scale with a lower footprint and do so much faster, significantly reducing the latency of time-to-scale. This positions Kubernetes as a natural fit for edge devices with resource constraints. And leverages containers, promoting efficiency and rapid deployment to disparate locations.
Efficient operations
Kubernetes excels in delivering simplified and predictable orchestration, reducing the complexity of deploying and scaling applications across geographically dispersed locations. Its ability to provide a unified management plane for diverse pipelines, over edge resources, simplifies the operation of what can be complex edge ecosystems.
Kubernetes not only solves many technical challenges associated with the edge, but it also lends itself to streamlined operations integrates into modern GitOps solutions. GitOps is a modern approach to software delivery and infrastructure management that leverages version control systems like Git as the source of truth for both application code and infrastructure configurations. In GitOps, changes to the system are made by updating Git repositories, and automated processes continuously synchronize the desired state of the system with the actual state, ensuring consistency and reducing manual intervention. This declarative and automated approach to operations improves efficiency, transparency, and the ability to maintain and scale complex systems.
One challenge of edge computing is consistent deployment and operations over heterogeneous infrastructure. When an organization rolls out hundreds or thousands of branches, it doesn’t happen overnight and the actual build can change over time. Kubernetes mitigates this problem by enforcing consistency in deployments across edge nodes, reducing the risk of configuration drift and enhancing reliability. It supports application portability and facilitates non-disruptive rolling updates and rollbacks, crucial for maintaining edge systems with minimal downtime.
Kubernetes brings with it built-in service discovery and easy to use load balancing mechanisms, that ensures edge devices can locate and communicate with services effectively, enhancing overall system performance and streamlining service chaining. Its extensibility and customizability enable tailoring to specific edge computing requirements, accommodating the diverse range of edge devices from small sensors and gateways to edge servers.
Last, Kubernetes is a great solution for brownfield solutions and as an intermediate step to solution on the step to full-blown Kubernetes as it can support both containers and legacy Virtual Machines (VMs), typically running legacy VMs faster than the legacy hypervisor-based systems they currently run on. So, there is no need to wait for your current applications to become containerized or even go end-of-support before they do so.
Other key technical advantages
Rapid scalability is essential in edge scenarios where workloads may vary significantly. Kubernetes addresses this need with automated scaling capabilities, allowing applications to seamlessly adapt to changing demands, ensuring optimal resource utilization.
Kubernetes promotes resource efficiency, a critical aspect in edge environments where resources are often limited. It optimizes resource allocation, minimizing wastage and ensuring efficient usage of available compute, memory, and storage.
With Kubernetes, operators designed for lightweight installation on edge devices, management becomes simplified. Kubernetes can also be integrated with observability tools, allowing monitoring of edge deployments, which is crucial for resource utilization and application performance insights.
Security is paramount in edge computing, particularly in sectors like healthcare and manufacturing. Highly functional Kubernetes distributions offer robust security features, including Multi-tenancy, Roles Based Access Control (RBAC), network policies, and pod security policies. This, in turn, can be coupled with state-of-the-art Kubernetes storage solutions that support data encryption art rest and in motion.
Not all Kubernetes solutions are created equal
While there may be more than one Kubernetes distribution that ticks all of the right boxes, it is important to remember it is not just what you do but how you do it. When evaluating a Kubernetes solution, it is important to consider the following:
How well does your solution support for mission critical and money making stateful applications including: content delivery, AI/ML, Security, finance and Industry 4.0 applications?
Does it provide application aware storage that allows one to unify data protection and disaster recovery of both applications and data into a single motion?
How does it optimize not just your platform footprint, but your data footprint?
Is it easy to use and does not require a PhD in Kubernetes itself as well as extensive command line operations. If it’s not easy to use, does it delay your deployments and lifecycle time to outcome?
Is it policy driven and support a set it and forget operations mode?
How securely, efficiently and easily does it implement resource sharing and unified operation of VMs and containers, eliminating operations and resource silos?
Is your vendor really willing to support both or is this just a bait and switch that can end up keeping you on your legacy VM solution longer?
Is your solution lightweight with full functionality?
Since many legacy Kubernetes distributions are enormous and lightweight solutions like K3s hamstring functionality and observability, find one that starts small, stays small and guarantees all of the same functionality at the edge that you enjoy at the core.
How does your vendor provide supporting operations and orchestration solutions that facilitate bare metal to service rollouts and lifecycle management?
https://symphony.rakuten.com/cloud/blogs/...enterprise-edge-computing
By Brooke Frischemeier Head of Product Management, Unified Cloud Rakuten Symphony
June 4, 2024
Over the history of data centers, we have seen trends come and go from on-premise, to client-server, to distributed systems and every possible term thrown in front of the word cloud, each of which is marketed as “the next big thing.” We have seen these changes and upgrades driven by numerous technological advances that created better business models, driven by facilities cost, Wide Area Networking (WAN) price-performance, mobile networking, chip-sets economy of scale, staff out-sourcing, data center repatriation and the list goes on and on. Fortunately, these models, along with their own lessons and gains are incremental, and it has led us to modern-day edge computing for many key verticals including retail, finance, AI/ML, Industry 4.0 and content delivery, just to name a few.
How enterprise edge changes the paradigm
Edge computing seeks to revolutionizing how data is analyzed, processed, stored, and acted upon. It represents a significant departure from the conventional cloud-centric approach by shifting compute resources closer to the data’s source, all the way to the "edge" of the network, closer to the user. This fundamental change to processing locality is ushering in a new era of responsiveness, efficiency, and scalability, with Kubernetes playing a key role in delivering this transformation.
At its core, edge computing is about decentralizing and distributing computing power and data processing, moving it from centralized cloud data centers, closer to the source of the action. In edge computing, data is processed and analyzed locally, without the need to send it over costly, slow WAN links, to central, far away data centers. This proximity to data sources significantly reduces latency and data bloat, ensuring real-time or near-real-time responses while reducing capital and operational expenses. It’s a win-win.
What are the key characteristics of edge computing?
Real-time processing: As we alluded to earlier, real-time processing is the hallmark of edge computing. It enables applications to make instant decisions based on data from sensors, cameras, and other sources, supporting smart cities, logistics, manufacturing and Internet of Things (IoT).
Reduce latency: Another challenge of the past has been latency, the bane of numerous applications over the years, and it is one of the most critical areas edge computing is helping to reduce. With edge computing we can drastically reduce round-trip response time and do so predictably. Many applications, such as augmented reality/virtual reality (AR/VR), autonomous vehicles, and remote experts require minimal latency to provide safe, seamless and immersive user experiences.
Data localization: With edge computing, data stays closer to its source, reducing the need for extensive data transfers and optimizing bandwidth usage. Data localization is particularly advantageous for remote or bandwidth-constrained locations. This in turn builds on lower latency/response times, by capitalizing local actions in industry, enterprises, content delivery and interactive networks.
Decentralized compute: Since edge computing decentralizes computational resources, distributing them across the network, it adds efficiency, fault tolerance and availability to the user experience. Centralized cloud services are vulnerable to network outages and latency fluctuations. Edge computing enhances reliability by allowing applications to continue functioning even when connectivity to the cloud is disrupted.
Scalability: Not all centralized clouds can be easily expanded upon, especially when you own and operate it yourself. Edge computing architectures are inherently scalable, as additional edge nodes and resources can be easily added, accommodating increased workloads and data volume.
Bandwidth optimization: Lastly, edge computing enables bandwidth optimization. Since data can be processed and reduced by edge applications, it reduces the strain on network bandwidth. Furthermore, it reduces the total amount of data that needs to be stored in a central data warehouse, reducing the costs of networking and storage resources.
In conclusion, edge computing is a transformative computing paradigm that brings data processing and decision-making closer to where it's needed most. With Kubernetes facilitating the orchestration and management of edge workloads, it is poised to play a pivotal role in enabling real-time, scalable, and responsive applications across a wide range of industries and use cases.
Kubernetes for the win!
Kubernetes has emerged as a compelling solution for edge computing due to several key factors that make it well-suited for managing distributed resources. Kubernetes comes with key set of advantages that align with the unique requirements and challenges posed by edge computing.
Kubernetes is an orchestration system designed to manage containerized workloads. Containers are lightweight, portable, and conducive to microservices architecture. This enables one to scale with a lower footprint and do so much faster, significantly reducing the latency of time-to-scale. This positions Kubernetes as a natural fit for edge devices with resource constraints. And leverages containers, promoting efficiency and rapid deployment to disparate locations.
Efficient operations
Kubernetes excels in delivering simplified and predictable orchestration, reducing the complexity of deploying and scaling applications across geographically dispersed locations. Its ability to provide a unified management plane for diverse pipelines, over edge resources, simplifies the operation of what can be complex edge ecosystems.
Kubernetes not only solves many technical challenges associated with the edge, but it also lends itself to streamlined operations integrates into modern GitOps solutions. GitOps is a modern approach to software delivery and infrastructure management that leverages version control systems like Git as the source of truth for both application code and infrastructure configurations. In GitOps, changes to the system are made by updating Git repositories, and automated processes continuously synchronize the desired state of the system with the actual state, ensuring consistency and reducing manual intervention. This declarative and automated approach to operations improves efficiency, transparency, and the ability to maintain and scale complex systems.
One challenge of edge computing is consistent deployment and operations over heterogeneous infrastructure. When an organization rolls out hundreds or thousands of branches, it doesn’t happen overnight and the actual build can change over time. Kubernetes mitigates this problem by enforcing consistency in deployments across edge nodes, reducing the risk of configuration drift and enhancing reliability. It supports application portability and facilitates non-disruptive rolling updates and rollbacks, crucial for maintaining edge systems with minimal downtime.
Kubernetes brings with it built-in service discovery and easy to use load balancing mechanisms, that ensures edge devices can locate and communicate with services effectively, enhancing overall system performance and streamlining service chaining. Its extensibility and customizability enable tailoring to specific edge computing requirements, accommodating the diverse range of edge devices from small sensors and gateways to edge servers.
Last, Kubernetes is a great solution for brownfield solutions and as an intermediate step to solution on the step to full-blown Kubernetes as it can support both containers and legacy Virtual Machines (VMs), typically running legacy VMs faster than the legacy hypervisor-based systems they currently run on. So, there is no need to wait for your current applications to become containerized or even go end-of-support before they do so.
Other key technical advantages
Rapid scalability is essential in edge scenarios where workloads may vary significantly. Kubernetes addresses this need with automated scaling capabilities, allowing applications to seamlessly adapt to changing demands, ensuring optimal resource utilization.
Kubernetes promotes resource efficiency, a critical aspect in edge environments where resources are often limited. It optimizes resource allocation, minimizing wastage and ensuring efficient usage of available compute, memory, and storage.
With Kubernetes, operators designed for lightweight installation on edge devices, management becomes simplified. Kubernetes can also be integrated with observability tools, allowing monitoring of edge deployments, which is crucial for resource utilization and application performance insights.
Security is paramount in edge computing, particularly in sectors like healthcare and manufacturing. Highly functional Kubernetes distributions offer robust security features, including Multi-tenancy, Roles Based Access Control (RBAC), network policies, and pod security policies. This, in turn, can be coupled with state-of-the-art Kubernetes storage solutions that support data encryption art rest and in motion.
Not all Kubernetes solutions are created equal
While there may be more than one Kubernetes distribution that ticks all of the right boxes, it is important to remember it is not just what you do but how you do it. When evaluating a Kubernetes solution, it is important to consider the following:
How well does your solution support for mission critical and money making stateful applications including: content delivery, AI/ML, Security, finance and Industry 4.0 applications?
Does it provide application aware storage that allows one to unify data protection and disaster recovery of both applications and data into a single motion?
How does it optimize not just your platform footprint, but your data footprint?
Is it easy to use and does not require a PhD in Kubernetes itself as well as extensive command line operations. If it’s not easy to use, does it delay your deployments and lifecycle time to outcome?
Is it policy driven and support a set it and forget operations mode?
How securely, efficiently and easily does it implement resource sharing and unified operation of VMs and containers, eliminating operations and resource silos? Is your vendor really willing to support both or is this just a bait and switch that can end up keeping you on your legacy VM solution longer?
Is your solution lightweight with full functionality? Since many legacy Kubernetes distributions are enormous and lightweight solutions like K3s hamstring functionality and observability, find one that starts small, stays small and guarantees all of the same functionality at the edge that you enjoy at the core.
How does your vendor provide supporting operations and orchestration solutions that facilitate bare metal to service rollouts and lifecycle management?
https://symphony.rakuten.com/cloud/blogs/...enterprise-edge-computing
|
Thread abonnieren
