In the world of long-haul network services and connectivity, enterprises seeking global WAN connectivity often benefit from a single service level agreement (SLA) with a global network services provider (GSP) that can provide support, comprehensive visibility & reporting.
But in an ever-shrinking pool of global network service providers, these companies lack the leverage to compel GSPs to meet and exceed service level agreements. Nowhere is this more evident than in the network performance of data-intensive applications such as voice and video where inadequate coverage in regional geographies is hampering service levels of the global wide area networks (WAN).
With enterprises demanding low-latency connections for their applications such as on-demand high-definition videoconferencing, IT procurement managers face important choices when engineering their global WAN architectures to meet this need.
Across a global WAN, the performance of delay-sensitive traffic such as voice and video may be within latency parameters (SLA) at first glance but the addition of local loop delay may render the application unusable. Application latency will normally be predicted by service providers as an average across their core network so the shorter the distances between the interconnects and hubs, the better the application performance in regional offices.
Emerging data-hungry applications depend not only on how much data can be transmitted but also how fast to ensure an optimal user experience. For example, many investment firms operate globally across multiple international trading exchanges where a single millisecond delay in a trade can mean millions of dollars in losses to the firm.
Another consideration with latency and WAN performance is validating data transferred at multiple locations using a ‘mirrored’ disk. When storing data remotely, the network user should see no difference between the co-located disk drive and the one at the remote location. But over time, too much delay between sending a signal to a remote site and ensuring the data packet arrives at its destination successfully will impede communication.
Latency can impact the global WAN performance in a variety of ways:
If your network has a high amount of latency, the bandwidth can decrease as result, so you won’t be able to send as much data. It’s worth noting that although your bandwidth won’t always be affected by your latency and may only last a few seconds, it can develop into a continuous problem.
IoT is a pervasive force in industry and the marketplace, affecting more and more businesses and vastly changing the way we view, receive, access, and use data. In a Futurum Research 2017 study, more than two-thirds reported that they were using the IoT in their operations. The effectiveness of IoT will depend on the responsiveness of a system or network to work effectively. High latency means delayed responsiveness and the inability of IoT devices and applications to function to their full capacity – or even as they need to.
Many companies are switching to Software-as-a-Service enterprise applications such as Asana, Skype, Slack, SharePoint, and Microsoft Office 365 to share and store data, and to collaborate on work internally and externally. If your network connection is plagued by high latency, you may experience lags or delays in your applications.
The fastest way to send data traffic across the global WAN is to reduce the overall path length. Having the shortest path will always minimise latency. Any excess length can add as much as two percent to the overall distance in minimising latency.
This is especially important for global WANs. For example, you may have a network service from a GSP for your existing offices but the company may want to expand its reach to remote sites in other regions. You need to determine if the current GSP can reach these new locations.
From the standpoint of governance and convenience, selecting your existing GSP simplifies contracts and keeps operational procedures similar. However some GSPs may not use their own network at the locations your company requires and may rely on partners to fill in areas that lack their own coverage.
How these partners are connected to the GSP’s primary network will impact the number of network hops and latency. This ultimately affects the performance of the workloads and access to applications.
This is where a network service provider’s capacity to provide PoPs in relevant geographies is often the critical determinant of the level of network latency and degraded application performance an enterprise can expect.
A PoP operates as a micro data centre that connects with upstream network providers to shorten the distance between two or more data transit points. PoPs in high Internet usage areas speed up responses to network queries over a global WAN. For example, with more PoPs in a localised network location, application requests are routed closer to the user than traditional data centres located in another country. Placing PoPs closer to end users as a way to reduce the number of hops supports content delivery networks for service providers while optimising use of network functions, virtualization, and software defined networking services.
Extending the PoP infrastructure into new regions, therefore, reduces latency for customers in those regions. For companies hosting new application services to their regional or remote offices, PoPs offer users more backup location options, reduce data transport costs, and ensure compliance with local data regulations.
While the breadth of services of a GSP can be valuable, sometimes depth or reach in a particular region is more useful. Adding a regional provider to your mix can often help you cut costs while achieving a business objective in a key market.
Some regions also pose technological hurdles that make it challenging for GSPs to deploy infrastructure locally. To overcome these limitations, regional service providers often build out multiple local PoPs and replicate the architecture quickly and efficiently within the local markets. These local PoPs are preconfigured with high performance connectivity features to optimise application performance and reduce latency.
Global providers may offer a compelling value proposition but if working in regional markets is critical to bringing your application to market, a regional service provider with multiple local PoPs can offer the right balance between performance benefits and cost savings. Selecting such a regional service provider as part of a global WAN architecture can increase agility, provide redundancy against outages, prevent vendor lock-in and increase your organisation’s reach and performance to help its business succeed.
As the leading RSP in Asia, Singtel1 offers the largest IPVPN network to connect business across Asia and to even the most remote cities. We provide integrated, fully-managed and secured networking service to help multinational corporations simplify their communications needs. Businesses only need to deal with a single contract and customer helpdesk on the back of one network to manage their communications across multiple countries.
For more information, please contact us here.
1 Singtel is the Market Leader in the 2018 IDC MarketScape: Asia/Pacific Next-Generation Telcos: Telecom Services 2018 Vendor Assessment