Slack, HipChat and Fleep are just a few of the new crop of real-time messaging…
Real-Time Voice Apps Require Deterministic, Intelligent Routing
Branch office network requirements are burgeoning and exceed the capabilities of conventional network architectures, a market transition primarily driven by the need for:
· Application scalability with increasing device-device latency/bandwidth communication
· Virtualization and hybrid-cloud multi-tier topologies
· Hyper-convergence forcing compute and storage to share a unifying network
Enterprise IT professionals and line-of-business end-users thus are looking for a new value proposition that results in an order of magnitude improvement in network performance and efficiency, scalability to multiple edge networks and thousands of virtual users, and fully sustained data bandwidth maintained as the network grows.
A fundamental problem associated with traditional WAN architectures is that of congestion. Managing congestion within multi-tiered, standards based networks is a key requirement to ensure high utilization of network resources and avoid the risk of:
· Networks that cannot scale in size to match application demands
· Slow and unpredictable network latency, reducing business responsiveness
· Unacceptably high cost of ownership due to over-provisioning
IT administrators are looking for more flexibility to solve issues of bandwidth allocation, network uptime and high availability, while driving down the cost of operations associated with multiple devices and carrier-class connectivity. One of the largest business drivers of demand on the edge-network branch is the widespread corporate initiative across all industries to drive digital transformation of standard business processes, which are way more reliant upon a mission critical infrastructure.
A University of Chicago study of more than 500 unplanned IT outages showed that networking failures are the second-highest cause of business impact to lost revenue, productivity drains and brand equity erosion. The applications that the study focused on were all digital transformation solutions running as SaaS or IaaS over traditional WAN topologies. Many of these apps also are designed to accommodate a mix of traffic, i.e., voice or video plus messaging, that often create havoc for static network routing. The apps studied included: chat (Google Hangout, Skype, WhatsApp); social media (Facebook, Instagram, Twitter); and video (YouTube, Netflix). Parenthetically, these are also the types of apps typically that can involve inappropriate use of network resources and even inadvertent propagation of network vulnerabilities that exploit rogue servers.
One area of traditional complexity is management of changes and updates to the legacy network equipment. A recent Gartner survey showed that 82% of network changes are still not yet automated. Given the diversity of updates constantly required – everything from OS version control to end-of-life issues to security patchwork – this is a significant gap in the high availability supply chain, in addition to an obvious drain on productivity, constant reactive or ‘firefight’ mode for network admins and a potential risk for contributing to brownouts or outright network disruption that thwarts business agility. Back to the University of Chicago study, more than half of human errors that caused a network outage were related to manual network updates and/or changes.
A next-generation WAN-Cloud solution ostensibly establishes a new paradigm in which packets are routed deterministically based on bandwidth suitability for the packet’s protocol, which is embedded in the header. This allows for low latency apps that are less business critical to be subordinated in favor of applications that touch customers and revenue. A good example would be a virtualized unified communications app deployed to accommodate both agents in a branch-office call center and remote independent agents linking in over a consumer broadband connection or wireless carrier network.
Another illustration of a real-time app with latency and VDI sensitivity would be emergency services and first-responder communications. The migration to Next Generation 911 (NG911) service is underway. With NG911, people can contact 911 via a text or video message from any device, as well as have caller-ID recognition. Under NG911, the location of cellular callers will be more accurate. The move to NG911 requires the migration to IP networks, rather than the traditional circuit-switched TDM networks that have been used for decades. 911 administrators need to evaluate adaptive-routing strategies that optimize network availability and resiliency.
Real-time apps like VoIP or 911 can be sensitive to packet loss, jitter, latency and congestion, and sometimes failed network links. Obviously, network outages, brownouts, dropped connections and poorly performing WAN links can turn VoIP or 911 into a busy signal — and that can seriously impact service delivery quality.
A failsafe SD-WAN (adaptive, deterministic intelligent routing) keeps VoIP and 911 services running smoothly. When an SD-WAN provides the last-mile connectivity to your branch office, data center or cloud provider, employees experience excellent quality for phone calls, virtual desktop sessions and other latency-sensitive applications.
A failsafe SD-WAN chooses the best path through the branch-office network, packet by packet at every instant. An SD-WAN solution proactively provide service-assured network quality, and can even switch dynamically between paths so fast that voice sessions don’t drop or degrade. The most advanced SD-WAN solutions, such as what @Talari provides, can selectively duplicate packets across the network, allowing session failover in a fraction of a second, fast enough to keep priority calls from dropping.