While you may associate web-scale networking with cloud giants like Facebook, Google, and Amazon, it’s not just an architecture for the large scale enterprises anymore. The industry has looked at data centers like theirs and asked the question: “What are they doing that we can mimic at a smaller scale?” Through analysis of the way these organizations ran, the term “web-scale” was born, referring specifically to the hyperscale website companies that have built private, efficient, and scalable cloud environments. Since then, it’s become a growing model for organizations to adopt in their journey toward evolving for the future.
The reason for this? The web-scale model offers many benefits for organizations who choose to adopt, including flexibility, resiliency and on-demand infrastructure. The open environment of web-scale makes it easy to standardize protocols, identify issues and create a unified stack that communicates efficiently. It allows organizations to combat the problems vendor lock-in poses by allowing organizations to customize based on needs and budget, and finally, it lets operators identify and fix issues quicker and deploy faster.
As the title suggests, this blog series will be all about how to network for the future, including tips, tricks, hacks and pointers on what the future of networking will look like, and how to navigate trends to find what’s best for your organization. I would be remiss if I did not start by thinking about the question of how this “futuristic” networking architecture of web-scale is impacting enterprises for the better. How are enterprises incorporating it, and where is it headed?
Embracing the futuristic mindset
According to Gartner, 40 percent of global enterprises will have a web-scale networking initiative by 2020. While this may seem like a high percentage, the truth is organizations have been slower to adopt web-scale principles than we would have imagined, despite the massive rewards of unparalleled scalability, efficiency and automation. Why is that?
For one, for a web-scale project to be successful, organizations first need to have the right mindset to implement it because it’s about a way of working as much as it is the technology. For example, one of the largest banks in the world is using Cumulus Linux to implement web-scale IT and it took them over four years to get to a position where they were able to fully commit to web-scale.
But, the mindset shift is really not as daunting as it might seem: in fact, your network team already has most or even all of the skills needed to execute a web-scale strategy, it’s just about learning the principles so they can think about their job in a new light. When an organization is built around web-scale principles, the IT team is responsible for improving capacity, and questions like the following become imperative to the mentality of IT operators:
How much compute do I have?
How fast can I store things?
How much can I store?
How often do things change?
What’s the horizon for growth?
What’s the strategy for failure recovery?
What processes can I make repeatable?
What can I automate?
The mindset of the IT team changes from a maintenance perspective to an architecture and design-forward perspective. This includes a focus on NetDevOps as a way to highly automate network infrastructure. By leveraging NetDevOps in an organization, the platform becomes highly flexible and scalable, removing much of the “red tape” in getting things done and embracing collaboration and transparency within teams.
The IT team has to have this sort of attitude focused on the future for all things web-scale to succeed in an organization, but while it may seem daunting to re-arrange IT team goals and encourage the team to embrace a new mindset, the payoff in the long run is massive.
Automation is key
When I talk to network operators today, they tell me they spend a lot of time doing network architecture, but don’t spend a lot of time focusing on automation. With web-scale, this flips: the IT team spends a lot of time on automation but not much time on operations itself.
Automation limits resource demand, mitigating errors and helping operators manage more switches. It allows rapid delivery of great applications and services to address business, predictable results due to decreased downtime and reduced mean time to resolution, and efficient management of nodes and devices, eliminating human error. From a networking perspective, the transition is all about knowing what’s going to happen before it happens: the IT team needs to be able to manage more capacity, more efficiently. They need a highly automated, self-checking supply chain that can drop capacity very cost effectively. With automation, the IT team will spend less time debugging and will have more time to manage more network devices.
Taking a leap for the future of IT
We have seen a good number of companies try to make the switch to web-scale incrementally. These companies make changes to adopt web-scale little by little, apprehensive about what it will look like for their company in the long run but willing to take only slight risks. It is great that these companies want to make the web-scale jump, but truthfully, these small changes will not create significant change and therefore will not allow them to reap the large rewards web-scale has to offer.
What we have seen work very effectively is communicating to the entire organization the importance of the web-scale switch to create a business of the future. Saying this shift is important for the business to sustain and succeed in the long run will make the team more inclined to adopt all at once, rather than incrementally. My advice would be to take the IT team of the organization aside and ask them to build web-scale infrastructure; then, the organization can start pushing new applications in web-scale, continue maintaining support for old applications in traditional infrastructure, and start migrating practices once it has established them.
The organization as a whole needs to be willing to uproot old practices and understand doing this is in favor of business advantage; risk-taking companies reap outsized rewards in their markets. The organization needs to be willing to take the calculated risk to achieve ultimate web-scale benefits and realize that by starting from the beginning, organizations can ensure teams are working together to think about web-scale design and manageability from end to end. Once you’ve established with your organization the new goals you’d like to achieve with web-scale, you can begin migrating every practice moving forward in the future to the “new” way of web-scale.
With the buy-in from your organization, you can build a web-scale data center with the benefits of standardization, predictable scale, lower TCO, business agility and customization. While it might take some work to get there at first, the major payoff for the business in the long run is undeniable.