With growing adoption of cloud infrastructures and Software as a Service (SaaS) in a hybrid-first reality, optimizing performance of enterprise networks has emerged as one of the top priorities across organizations. Today, nearly every enterprise uses a range of wired and wireless connectivity technologies, as well as multiple protocols to connect a growing number of endpoints and services. It’s vital they do everything they can to optimize performance.
In the early 2000s, most enterprises deployed appliances that helped them optimize the performance of their wide area networks (WANs), largely orchestrating traffic between the office and the data center. These appliances were built with the assumption that:
The network typically leveraged private links like MPLS, FR, and T1 for critical traffic, using WAN optimization appliances to solve key challenges:
Traditional WAN operations, while they were prevalent 20 years ago, fall short in the context of modern-day enterprise networking needs. Last-mile bandwidth is significantly cheaper now, mission-critical applications reside in multiple clouds and in the data center, and endpoints are distributed around the globe today. What’s more, internet bandwidth is much cheaper and can be provisioned much faster as compared to private links.
Plus, WAN operations may not even be possible in scenarios where enterprise applications reside in the public cloud. And even when it is, enterprises look to cut network service costs by provisioning support at a single location instead of appointing network engineers for satellite locations like warehouses, branches, retail outlets, satellite data centers, and so on.
Clearly, network optimization carries a different meaning today. Here’s what’s changed:
The software-defined wide area network – or SD-WAN – is a low-touch virtual architecture that is leveraged to orchestrate network traffic across a variety of connections, while ensuring a high quality of service for end users.
While SD-WAN is a networking paradigm that is rapidly replacing traditional WAN architectures that depend strongly on private links, its impact on network optimization capabilities are wide-ranging.
Modern enterprises not only use a variety of applications, they also adopt new applications every few days while retiring old ones. Some of these generate and attract real-time traffic, others generate transactional traffic, and yet others require bulk transfers. What’s more, some of these applications may not be critical and still encrypt data in transit.
To optimize traffic in such a scenario, the solution must be able to assess the criticality of the application through built-in techniques, and recognize applications through Layer 2 to Layer 7 attributes. SD-WAN solutions will typically not only come with these built-in capabilities, but also apply intelligently configured defaults at the time of implementation. They are also able to determine which applications need multipath connectivity and which can be provisioned by a single channel.
While broadband links may not be as reliable as private links, they offer much higher bandwidth at lower costs. This can be exploited to provision multiple links for failover and congestion handling. However, to optimize traffic across these links, IT must deploy a solution that detects QoS drop and failure within milliseconds and steers traffic to an alternative link quickly to keep the user experience intact.
This is precisely what SD-WAN solutions do – they continuously monitor the end user experience based on latency and other metrics, and scan for congestion to efficiently route traffic along the most optimal link. Some solutions are also able to detect asymmetry and reset connections to restore the application traffic to an optimal state without manual intervention.
In hybrid work models, enterprises have rapidly adopted real-time collaboration applications. The efficacy of such applications is strongly impacted by small jitters, and lossless transmission is crucial, even if at the cost of generating redundant traffic. Moreover, the participants involved in a scenario may be distributed around the globe.
SD-WAN solutions leverage techniques like forward error connection (FEC) to empower high-fidelity collaboration via such applications. They also use other algorithms to eliminate redundant packets and reorder packets to prevent applications from sensing a lossy scenario. With these strategies, SD-WAN solutions are able to achieve better performance than in scenarios where a dedicated MPLS link is used.
WAN solutions may still be relevant for middle-mile issues that impact network performance (like congestion), and approaches like compression, protocol acceleration, and de-duplication will still be relevant in these scenarios. However, it may not be possible to deploy WAN operations appliances in conditions where SaaS or PaaS traffic is involved.
With SD-WAN, IT can provision these capabilities near the point of requirement with the use of a hypervisor. In fact, network administrators are able to deploy any networking capability within seconds at any point in the network, enabling them to optimize network traffic.
Finally, some SD-WAN solutions will further optimize the overall orchestration strategy with universal de-duplication and techniques like stream compression to achieve even better networking and QoS metrics.
Hybrid networks are becoming increasingly difficult to manage with traditional approaches, and legacy WAN operations techniques are rapidly becoming irrelevant, given the needs of modern-day enterprises. SD-WAN solutions, however, enable network engineers to keep their networks functioning at highly optimal levels by default. They empower low-touch network operations and pave the way for incorporating security into network orchestration considerations with the secured access service edge (SASE) paradigm.
If you’re looking to optimize the hybrid work environment for your business, you can turn to Zones for support. Connect with a Zones Account Manager today.