As enterprises embrace digital transformation, particularly the rapid adoption of SaaS applications and the migration of enterprise workloads to IaaS, it has become critical for businesses to view their IT investments as a way to maintain a competitive edge.
Achieving the highest cloud application performance is no longer an option but a present day imperative. Any additional latency encountered while accessing applications or workloads in the cloud directly impacts the quality of the user experience and business productivity. Dissatisfied users can mean lost revenue, customers and employees, all of which are critical from a business standpoint.
Overcome latency issues
To mitigate latency challenges, enterprises have traditionally signed SLA agreements with their service providers and typically subscribed to private MPLS transport services. Provisioning a private leased line connection such as MPLS can be an expensive proposition.
Unlike traditional data centre hosted applications, however, cloud applications and workloads have altered network traffic patterns. To deliver the highest quality of experience, IT must transform the network.
Traditionally, remote users at branch sites would connect to data centre applications using an MPLS connection. But the world has changed. Today, SaaS applications like Salesforce, Box and Zoom, and public IaaS instances like AWS, Azure and Google Cloud Platform, are causing enterprises to rethink their networks.
With the majority of workloads running in the cloud today, new challenges are emerging. For instance, using an MPLS connection to backhaul all cloud-destined branch traffic to the on-premise data centre no longer makes sense. It results in network congestion and adds latency, not to mention unnecessarily consumes expensive MPLS bandwidth.
Latency is the number one enemy of delivering a high-quality user experience. The more latency that’s introduced into the network path between the user and the application, the more the user experience degrades.
To counter this challenge, enterprises are deploying an SD-WAN platform for its ability to enable secure local internet breakout, offer sub-second failover between underlay links and use any combination of transport services including broadband internet, to automatically direct cloud-destined branch traffic directly to the cloud.
Enterprises gain the ability to reliably and securely connect branch users to cloud applications and services, even when an underlying link experiences congestion, brownout or outage.
Optimise cloud application traffic
To counter the unpredictable nature of the internet, advanced path conditioning corrects lost and out-of-order packets by reconstructing or resequencing them at the receiving end. This overcomes the impact of packet loss typically observed on internet connections.
An advanced SD-WAN platform should automatically identify and classify applications on the first transmitted packet. Then it should select the best performing internet connection to connect users to a cloud application based on the link performance when two or more internet services have been provisioned. This improves the path selection intelligence for organisations that breakout internet traffic locally to deliver the optimal cloud application performance.
Often organisations provision two or more WAN links from remote branch sites to increase network and application availability and performance. To optimise the use of provisioned WAN internet links, performance can be monitored in real-time by continuously measuring packet loss, jitter, latency and mean opinion score (MOS).
Statistical learning can also be used to dynamically determine the optimal link for breaking out traffic, which maintains peak application performance. This optimises internet breakout traffic to deliver the highest cloud application performance.
Highest quality of experience
Selecting the best path to direct packets eliminates any additional latency experienced by cloud applications and workloads. This delivers the highest quality of experience to users and results in happy customers, prospects and employees.
In addition, brownout or blackout conditions can automatically be addressed on any link. For instance, if a branch is served by one MPLS connection and one internet connection, the connection quality can be confirmed before sending any packets over the internet. If for some reason, the internet connection is experiencing loss, latency, jitter or MOS greater than a pre-configured threshold, the MPLS connection will automatically be selected to send packets.
Tunnel bonding also supports packet-based load sharing and higher application availability. Multiple underlay links can be bonded into a single business intent overlay, making efficient use of all available bandwidth. This is achieved using techniques like Forward Error Correction (FEC) and Packet Order Correction (POC).
All of these factors ensure that no matter what happens, enterprise applications always operate at peak performance, regardless of whether they are hosted in the data centre or in the cloud. This dramatically improves business productivity, while enabling businesses to increase efficiency. All of this is crucial in meeting the needs of the modern cloud-first enterprise.
About the author
Dean Vaughan is vice president for Asia Pacific and Japan for Silver Peak. He is responsible for enterprise and service provider sales, including go-to-market and channel strategies to drive partner engagement and customer acquisition across the region. For more information, visit: https://www.silver-peak.com/