Laws of Network Physics Hamper Rush to the Cloud

When the COVID-19 pandemic first hit, the assumption was employees would be working from home for a few weeks to combat the spread of the virus. Many IT teams viewed the pandemic as a temporary event that would require the bulk of employees to work from home in much the same way a blizzard paralyzes a city for a few days.

Those “snowdays” where employees need occasional remote access have now become the new normal. In fact, five months after the first orders to work for home were issued, it looks like most employees will be working remotely either completely, or at least a lot more often, well into 2021.

As a result, organizations of all sizes have accelerated the rate at which application workloads are moving into the cloud. Those workloads are typically accessed via a virtual private network (VPN) or over a software-defined wide area network (SD-WAN). The issue that many of those organizations are now contending with is that the laws of physics have not been suspended. Applications remotely accessed over a network are going to be slower than applications accessed over a local area network in an office.

More challenging still, employees are either accessing cloud applications directly over a public Internet connection from home or they are using a VPN to access cloud applications via a corporate network through which network traffic is backhauled. Thanks to the laws of physics, all that additional latency tends to have a negative impact on application experience. In fact, employees residing in one geographic region can easily be having a completely different application experience than others.

Application performance hindered most by ISPs

As it turns out, the entity that delivers the networking services that enables cloud computing can have a profound impact on those application experiences. A report published by ThousandEyes, a provider of Internet monitoring services that is now an arm of Cisco Systems, finds that globally Internet service providers (ISPs) between January and July experienced more than roughly 4,500 outages. That compares to about 400 outages experienced by cloud service providers. Relative to total outages for the period, more than 80% occurred within ISP networks compared to less than 10% within cloud provider networks.

Most cloud service providers have deployed their networks more recently than ISPs, so they currently enjoy advantages in terms of performance and resiliency, says Angelique Medina, director of product marketing at ThousandEyes.

“There’s a significant difference,” says Medina.

Sooner or later ISPs will need to update legacy network infrastructure to better service cloud applications if they hope to avoid being usurped by cloud service providers, notes Medina.

IT leaders driving the transition to the cloud clearly need to investigate the quality of the network services being relied on to deliver any cloud application experience. All the advanced networking software and hardware deployed at the edge of any network isn’t going to make much of a difference if the core – the wide area network – is archaic. No matter who might actually be responsible for the delivery of a network service, it’s ultimately the IT leader that is going to get blamed for any and all suboptimal cloud application experiences. As such, network performance, and by extension architecture, in the age of COVID-19 now matters more than ever.