Version française prochainement disponible
By: Robert DeWeese
Say you lead a small but highly competent networking team – working for a small but prestigious health care provider or bank. A few years ago, your team got a new deployments system up and running. It’s worked well so far, but now your organization is scaling up and the cracks are starting to show.
The manual side of cloud maintenance was always an annoyance for your team, but it was never prohibitive—until now. Today, cloud maintenance has become so labor intensive that you worry it’s only a matter of time until a mistake occurs, and a virtual door is left open to exposure or breach.
Now: scenario two.
You lead networking for a multinational hotel or car rental brand. You’ve spent the last few years migrating to a software-defined wide-area network (SD-WAN) network, supported by a private multi-protocol label switching (MPLS) network. You have software-defined routers at each of your 3,000 branch locations, each requiring $1,000 a month in licensing fees. On top of this hefty monthly bill, whenever you need to add a new branch to the network—granting access, authenticating, and other tasks—you deal with delays of up to 90 days.
Clearly, the networks in both these scenarios need a cloud makeover—a new roadmap for reducing monthly operational costs, better securing data, and reducing the chance of outages, exposures, and breaches.
I’ll suggest how to approach the task, but let’s first cover why these scenarios are so common. In other words, why have cloud networks become so messy in recent years?
Starting again from square one is not an option. Instead, enterprises need to re-evaluate how their clouds are managed, controlled, and secured.
Growth without architects
Even with advanced container orchestration system tools, and even when working within databases, spinning up a cloud environment always involves networking. Over the last 10 years of migrating to cloud, I’ve watched as many networks have been built without the guidance of a cloud network architect. As a result, these environments have grown incredibly complex.
Starting again from square one is, of course, not an option. Instead, enterprises need to re-evaluate how their clouds are managed, controlled, and secured. For many teams, it means beginning with the basics—in other words, the budget. Messy clouds frequently equate to gross overspending, so starting with an assessment of quarterly spend to ensure operational costs are optimized is often the first step towards a healthier cloud.
Keeping pace with hyperscalers
Staying up to date with new cloud products and features is equally crucial. Hyperscalers keep cloud technology moving forward at a ferocious clip. Your team must be able to determine which elements of your own infrastructure no longer carry weight, and which new features and updates can help maximize your cloud’s performance and efficiency. This step is best fostered by a corporate culture committed to upskilling, which allows your network admins to evolve into engineers, and your engineers into architects.
While this all might sound simple enough, navigating a more active approach to cloud—and ultimately, the constant cycle of evaluation and re-evaluation—requires a unique set of tools and support that may not yet be standard practice for your organization.
Operating on a secure cloud is not only vital to basic operations, but also for working more efficiently and effectively—and for staying out of the news.
Finding the right solution
For many smaller teams, the idea of change is often a fraught one. In my experience, I’ve found many teams feel hesitant to automate because they perceive automation as a threat to job security. When it comes to the cloud, however, automation really means more control over the environment, more security, and therefore, more peace of mind for everyone involved.
For example, with the type of data that a healthcare provider or bank handle daily, operating on a secure cloud is not only vital to basic operations, but also for working more efficiently and effectively, and—importantly—for staying out of the news.
Your cloud environment should be what allows your organization to grow, not what holds it back from scaling up. That’s why the teams in these scenarios would benefit from outsourcing the maintenance and monitoring of their cloud environments to a managed, backbone service like Cloud Flare or Aviatrix Solutions. Through services like these, they can continue to grow their operations while ensuring their multi-cloud, multi-SaaS environments remain as secure as possible.
Burden of sunk costs
Going back to the second scenario with the multinational organization: hesitancy to invest in a cloud makeover might have less to do with fear of automation and more with the burden of sunk costs and time.
Perhaps surprisingly, the right solution for the smaller team might also be the right solution for the multinational one. Once the larger operation is riding the cloud backbone by way of a managed backbone service, it can break free of the traditional MPLS and SD-WAN space and its expensive licensing and hardware footprint.
Instead, all the company needs are some inexpensive routers to place at each of its branches to connect the network to the closest peering point. This step not only reduces attack vectors, but also offers the company a simplified policy-control mechanism for what has become very complex inter-cloud, inter-SaaS, inter-branch communication.
Designing for growth
Now, say the car rental brand wants to grow revenue by monetizing its data. To do so, the company invests in a fleet of smart cars to not only make customers’ rental experiences more enjoyable, but also generate heaps of new data.
Harvesting the data will involve a lot of complex networking. The company would need to bring in feeds from the smart cars, send that data to the closest cloud provider, dump it in a data lake, and further manipulate it before that data can be harvested and sold.
That’s where a tool like Cloud WAN comes in.
Cloud WAN, a distributed computing architecture from AWS, enables cloud admins to create overlays. In turn, the overlays enable the admins to work more efficiently. They can spin up an overlay for a client that allows access to dump data into the data lake. A separate overlay can then enable communication from the data lake to the smart capabilities accounts before feeding it back to the smart car sensor and, ultimately, the smart car app.
A cloud makeover is not, by any means, a small step to take. It requires time, patience, and the willingness and participation of stakeholders up and down the enterprise.
But it’s worth the effort—the security, efficiency, and potential for growth that a cloud makeover can bring is nothing short of transformative. From my perspective, it’s the only way forward in this ever-evolving cloud landscape.
Robert DeWeese is Director of Cloud Networking at Kyndryl.