While countries around the world struggle to get a handle on the COVID-19 pandemic, engineers are watching and tweaking their networks to accommodate sudden growths of traffic, as students move to online classrooms, their parents work from home, and seemingly everyone turns to video conferencing happy hours and streaming entertainments for the stay-in evenings.
At first, the traffic changes would be small ones, as the internet has long been engineered for the crush of streaming services such as Netflix. But cracks are appearing: network intelligence firm ThousandEyes, for instance, found that ISP and cloud service outages have risen 42% over the past six weeks, as traffic has spiked by 20% of Verizon’s networks and Vodaphone by 50%. As a result, enterprise service providers have had to make some tweaks as their customers start dialing in from home. For example, Microsoft, to meet the increased loads on its Teams collaboration platform had to adjust for “non-essential capabilities to help manage service quality,” reported SDxCentral.
Everyone staying at home may very well bring large-scale, and perhaps, permanent changes to both networking topologies and computing platforms. This week on The New Stack we will be surveying the service and tool providers to learn more about what they are seeing on the front lines, and what changes they see afoot. First up: We spoke Apurva Joshi, vice president of products at DigitalOcean, what about this cloud services giant is seeing in terms of impact, and the success it has had in keeping its customers happy.
What traffic changes are DigitalOcean and its customers seeing due to behaviors around COVID-19?
We’re seeing early indications, especially from customers in the video streaming and gaming workloads, that usage is rising considerably. Notably, customers are spinning up more compute nodes to keep up with the growing demand, especially as the percentage of people working from home, and staying home in general, has spiked so dramatically.
What changes, if any, has DigitalOcean made to respond?
We usually plan our capacity six months in advance, but the current, unexpected situation has forced us to fast-forward our usual process. For example, we’ve added more capacity to our data centers than usual to make sure our customers are not facing interruptions.
DigitalOcean also recently launched the Hub for Good, which is intended to support our community of innovators and technologists who are leveraging their skills to create tools and resources focused on the COVID-19 pandemic. As a result, we’re committing $100,000 in infrastructure credits to new, not-for-profit projects related to COVID-19 as well as $50,000 in cash donations to the COVID-19 Relief Fund. This is a shift away from our core product offerings but at the same time it is true to our mission of supporting our developer community — and the community at large — with the tools they can use to change the world.
Are there spikes in certain workloads? Do other workloads need to be scaled down?
As noted above, the most notable increase we’ve seen so far is from customers in the video streaming and gaming workloads. Another is DigitalOcean’s Managed Kubernetes and Bandwidth usage, which has seen about 20-25% spike in the last few weeks. Overall, there has been a change in the network and scale characteristics across the board, and we are monitoring the situation daily so that we can best serve our customers and stay ahead of the evolving crisis.
Are you seeing scalability issues in end-user apps?
Elasticity is built into public cloud infrastructure, which enables end-users to scale their workloads up or down with few clicks. DigitalOcean serves SMBs [small and medium-sized businesses] and individual developers with a global presence of 12 data centers around the world. This combination of our infrastructure scale, advanced capacity planning and targeted customer mix has allowed our end-users to scale their applications without any issues. We continue to monitor the situation on a daily basis and serve our customers efficiently.
Are you seeing network saturation, or a need for network rebalancing?
While we are experiencing increased network utilization, we haven’t been hitting any limits on the network side. We continue to rebalance and augment our compute workload with our popular High CPU, General Purpose and Memory Optimized droplets to meet demands from new and existing customers.
Are you seeing issues with staffing — are places suffering from COVID-related absences?
As an organization that is 70% remote, we already have the processes and systems in place to support our now fully remote workforce, Additionally, our suppliers, international shippers, third party remote hands, and internal staff are currently operating as expected and we have no immediate concerns around our infrastructure capacity.
Do you see any changes that could possibly be long-term in effect (a “new norm” if you will)?
Right now, the focus is justifiably on navigating through the current crisis and ensuring our collective welfare. In time, this may be seen as a watershed moment that drives an increase in remote work, and heavier investment in cloud technologies that enable such environments. Events like these will reshape our society in lasting ways, from how we work and collaborate, to the level of security and surveillance we’re accustomed to, and even the new products we build to adjust to the new behaviors of our developers and SMB community.
Before the crisis, we all expected that the remote work culture will be prominent in the coming five to six years. These events simply fast-forwarded those timelines. Digital Transformation will be a top priority in every business conversation. The urgency for SMBs to have an online presence will continue to grow exponentially.
Our simplicity of products and pricing, especially bandwidth costs and community support will continue to be our unique strength. The opportunity ahead of DigitalOcean to serve our customer segment will be bigger than ever before.