×
Register

Already have an account?

By Signing up for Swiftnlift, you agree to the Terms of Service and Privacy Policy of the platform

Industry Viewpoint

Pi Data Centers
no image

Will the Demand for Hyperscale Data Centers Fall as Edge Data Centers Proliferate?

Debmalya Dey Roy (Deb) Vice President, Pi Data Centers


When we hear the term "hyperscale," we immediately think of something massive/colossal. However, in the data centre industry, hyperscale does not simply refer to the amount of space available, but also to how more equipment, higher power capacity, network abundance, near zero downtime, and process maturity can be accommodated and served to clients with mission critical operations. Hyperscale data centres are well-known for their ability to provide reliable, scalable infrastructure with measurable SLAs. Because this type of data centre is typically operated by large service providers, a hyperscale data centre is known for its exponential scalability in a short period of time rather than its size.


There is no doubt that the edge is the way forward for enabling data ease of access and lowering latency. With next-generation IoT and virtual reality applications taking centre stage, hyperscalers will play a more strategic role in terms of capacity expansion and reach. While discussing hyperscale enterprises across industries, most of us tend to overlook the fact that these are not just big names, but also the platforms that meet the needs of millions of inquisitive customers who rely on hyperscalers to solve their myriad problems. This boils down to providing bits as numerous smartphones, tablets, desktops, and so on in industries such as SaaS, content-based OTT players, and creators of digital assets.


Given that the volume of these exists in direct correlation to population density, it's no surprise that hyperscale data centre providers are constructing availability zones in more population centres. Beyond Tier 1 cities, global public cloud players are expanding their "Local Zones." OTT and other real-time data delivery platforms are also expanding their footprints and gaining a foothold in Tier 2 and Tier 3 markets. That is where the presence of edge data centres would be required. In tier 1 and tier 2 markets, we are seeing a shift from hyperscale monolithic data centre providers to hyperscale data centres with edge child locations (hub and spoke). The reason for this is that these cities and their surrounding areas have a high concentration of digital content consumers.


The Impact of Hyperscalers on Edge Strategy


It is clear that hyperscale providers and the customers they serve are expanding beyond a few regional availability zones to deliver digital services to a much larger consumer base across a broader geography. Moving to the Edge and delivering will be the focus of the next five to ten years. However, it is unclear what types of applications and use cases will necessitate such a highly geographically distributed footprint. In addition, how would application developers and content producers manage, scale, and allocate geographic resources in a fragmented construct?


Regardless, as cloud and technology providers shift to a local-zone edge strategy, they will seek to address three logistical issues:


1. Metropolitan Diversity – In the same way that hyperscalers built their initial regional availability zones in three-tripod groups for resiliency, they will look to do the same in the metro or local availability zones into which they expand. This necessitates the establishment of at least three data centres in each market. In contrast to single-location wholesale data centres, it would necessitate the data centre partnering with multiple facilities.


2. Robust Network Connectivity – Local availability zones require ample fibre networks and carrier-neutral interconnection hubs to connect the nodes in that metro to one another, as well as to connect the local zone to those in other regions. Hyperscalers will look for metros with a diverse set of neutral interconnect hubs and fibre paths in and out of the metro. Hyperscalers are likely to favour a multi-location model.


3. Competitive Speed to Market – Traditionally, hyperscalers would rely on a combination of their hub hyperscale facilities and spoke edge data centre locations to meet demand. As hyperscale data centres, they would be able to support local zone edge deployments with 5MW+ power and pull dark fibres to the edge connecting their hub location. As a result, existing enterprise data centres with significant capacity in tier 1 and tier 2 metros will be an appealing proposition.


We can clearly see Hyperscale Data Centers and Edge Data Centers coexisting and serving the industry's larger need of bringing data closer to the last mile user without sacrificing service quality or service availability.