Beyond Latency - Reliability, Security, and Operational Simplicity - the Other Key Drivers of Distributed Computing
April 7, 2021
Latency is often considered the primary driver of the edge computing movement. However, there are other factors, beyond speed, that are pushing this architectural paradigm shift. In this post, we’ll look at how latency, reliability, and security are accelerating edge computing innovation and adoption, and why the demand for operational simplicity is a parallel thread to total Edge application delivery.
Latency: The need for speed
Applications today are far different than they were ten years ago. Adding to core web application and gaming needs, even in the last two years, we’ve seen dramatic innovations in technology being driven by emerging use cases such as autonomous vehicles, smart cities and homes, real-time communications, and many more. What all of these use cases have in common is the need for a move from human to machine speed.
Edge computing enables developers to design applications that capture, store, and process data closer to where it’s generated, delivering dramatic performance improvements, alongside bandwidth savings and other benefits.
However, as we’ll discover in the next sections, reduced latency isn’t the only driver of the move to distributed computing architectures.
Reliability: Designing for failure
Before optimizing for speed, application architects are faced with the challenge of designing systems that enable workloads to perform their intended functions accurately and consistently when expected. Reliability is a cornerstone of application design and one of the primary considerations leading developers to the edge.
Each year, media sources chronicle all the major cloud outages that took place over the course of the year, e.g. 5 Cloud Outages That Shook the World in 2020. In many of these instances, core services were impacted, often going offline for an extended period of time and taking down many highly trafficked applications in their wake. This centralized blast radius inevitably makes application architects question deployment models.
The only guarantee when it comes to computing and networking infrastructure is that there will be failures, and downtime translates to money lost. When all operations are concentrated in a centralized system, all it takes is one point of failure to disrupt or bring down the entire system.
Designing fault tolerant, resilient systems is no small feat. Application architects are increasingly looking to hybrid, multi-cloud, and cloud-edge models of deployment in order to mitigate the risk of degraded service and outages. Building redundancy into systems allows for failover to healthy systems when something goes wrong, minimizing the impact of incidents.
Security: Compliance and threat intelligence
As the pace of technology innovation continues to accelerate, so too does the growing sophistication of threats, targeting, and attacks. On top of that, increased regulatory requirements are forcing organizations to rethink how, where, and when they process and store data.
Edge computing topologies are extending flexibility and granular control to application architects to meet compliance requirements, while also mitigating the impact of potential security breaches. Processing data closer to end devices allows for earlier threat detection and mitigation, before attack agents are able to penetrate mission-critical operations.
Furthermore, the amount of data needing to traverse the network is significantly reduced, minimizing the attack surface and allowing application architects to focus on protecting the most vulnerable vectors. When an attack occurs, the distributed nature of edge computing infrastructure enables security protocols that are designed to seal off compromised services without shutting everything down.
Historically, the notion of moving elements of an application to the Edge has been associated with increased complexity. An additional delivery plan to manage means an additional deployment and diagnostic plan to consider for operational management. However, latency, reliability, and scalability benefits of moving parts of the application to the Edge have outweighed the downside of increased operational complexity.
Edge developments over the last few years have now turned this paradigm on its head. New Edge as a Service (EaaS) solutions present the opportunity for engineers to place the entire application at the Edge and deliver it from one plane, simplifying the operational and deployment model.
Compared with the legacy cloud-edge hybrid model, moving 100% to Edge is bringing operational simplicity.
As the global population depends more and more on interconnectivity of data and applications, deployment models need to keep up to ensure fast, reliable and secure experiences for end users.
There is no one size fits all when it comes to architecting optimized systems, and there are often trade offs based on prioritization of objectives. When asking what’s most important for a particular organization or application, it’s important to evaluate all factors alongside each other. For example, when does resilience matter most? When is reliability most important? Latency? Security? There is no single answer to any of these questions, and the answers are always evolving.
The extension of cloud to edge is giving application architects more flexibility in being able to address all of the above questions in more granular terms. However, with increased granularity comes greater complexity. Edge as a Service (EaaS) is helping to simplify this edge-cloud puzzle.
Want to learn more? Start customizing your edge solution today.