Data Center Architecture Lessons From Isaac Newton
Sir Isaac Newton remains our favorite source for axiomatic laws of physics, despite giving us the language of calculus. Particularly relevant for today’s discussion is Newton's third law as formally stated: “For every action, there is an equal and opposite reaction.”
In the cosmology of the data center, this existentially proves itself in the network whenever there are significant changes in application infrastructure and architectures. As evidence, consider the reaction to first, virtualization, and now, containerization, APIs, and microservice architectures.
These changes, while improving speed and agility of application development and delivery, have created greater mass in the data center, essentially changing the center of gravity and pulling many network services toward it. Application-focused services like load balancing and even web application security have been pulled toward the development environment as scale and security have become a necessary component of application architectures.
The result is a growing application network that is separate from “the network” in which developers and operations (DevOps) tends to provision, manage, and deploy not just the services and apps, but the network services necessary to support them.
But the core network, and the need for it, has not diminished. Indeed, it has grown to epic importance to the business, as the core network becomes the primary lifeline through which all data flows, both inbound and out. Should that lifeline falter, or slow, business will be impaired. Productivity will plummet, and profit will plunge. Brand reputation will suffer and woe unto those who failed to maintain the performance, stability, and security of the network upon which business today ultimately relies.
This is the reality of today’s data centers; the inevitable Newtonian reaction to the changes wrought by emerging technologies and architectures in how we develop and deliver applications.
The core network (pictured above as the N-S network) is for business. Inbound and out, it's critical to the success and growth of every business, no matter its vertical industry. The app network -- the E-W network -- is for apps. Whether APIs or microservices, monoliths or databases, it's critical to the scale and architectural integration of the hundreds of applications, services, and data sources that feed the rest of the business and enable consumers to interact with business.
Both the core network and the app network are important pieces of the puzzle, but both have priorities that cannot be met by the same solutions anymore. Business needs agility from its applications, but it requires stability from its network. You cannot have one without the other. The scale required of the core network is business class; in the app network, it is per-application class.
To achieve the kind of business scale appropriate to managing the volume generated by inbound and outbound use, while simultaneously fending off the inevitable attack, requires greater capacity and scale -- the kind of scale associated with hardware. Routing, switching, security, and delivery at the business level is hardware-based, as it has been since scale and growth drove the market to that realization.
Achieving the agility and speed necessary in the app network requires software -- virtual or containerized -- with APIs and programmatic methods of integration into the orchestration engines driving the build and release process. These components must also be scalable, but not of the same magnitude as the core network. Software solutions rule in this growing division in the data center.
As the impact from emerging markets like Internet of Things and more changes in applications -- think Apple’s mandate to move all apps to HTTPS by 2017-- the core network will experience additional pressure in the form of increased demand for more traffic and more processing without any disruption of stability or speed. Hardware solutions continue to be the most efficient means of achieving both in the N-S network.
Software solutions will no doubt be the future of the E-W network, providing agile, disposable infrastructure for rapidly changing application and service architectures. But both networks are equally important, and organizations can alleviate some of the pressures by adopting a clearer, two-tier network approach to data center architectures that recognizes the needs of each and answers them with the appropriate solutions.
Recommended For You
Innovative companies are leveraging open frameworks such as composable infrastructure to forge a path toward making fabric-based infrastructures a reality.
Anyone deploying NVMe-oF should examine their existing infrastructure investments, IT roadmaps, and expected workloads to choose the right flavor of NVMe-oF.
By enabling up to 30 times faster deployment than firewalls, and simplified management, segmentation reduces CAPEX and OPEX compared to traditional techniques.
DevOps principals have influenced how infrastructure is managed at scale using IaC strategies.
To meet increasing consumer and enterprise demand for data (especially from video streaming and IoT devices), data centers must upgrade their infrastructure.