An edge data center is a data center located close to the users, devices or networks it serves, rather than in a large central hub. Edge data centers are usually smaller than core facilities and are often found in regional cities, at telecom facilities or near cell towers. By putting compute, storage and network connections nearer to where data is created and used, they reduce latency, cut the cost of carrying traffic long distances and can help keep data within a region. Many operate as colocation sites, renting space, power and connectivity to businesses and service providers.
At a glance
- A facility type defined by location close to users, not by a specific size or design.
- Usually smaller than core data centers, from regional buildings to modular units at telecom or tower sites.
- Used for latency-sensitive applications, content delivery, and local processing of large data volumes.
- Reliability, connectivity options and services vary more widely than at large facilities.
What problem it solves
Large data centers tend to cluster in major hubs where land, power and network connections are plentiful. That works well for most workloads, but users and devices far from those hubs see higher latency, and their requests have to travel over long-haul networks. For real-time applications such as interactive video, gaming, industrial control or analysis of camera feeds, those extra milliseconds and the cost of hauling large data volumes can matter.
Edge data centers place infrastructure closer to the demand. A business can serve users in a secondary market from a local facility instead of a distant hub, process data from many sites in a nearby location before sending summaries to the cloud, or reach local networks more directly.
How it works
Location. Edge facilities sit in regional cities, suburban areas, at carrier central offices or cable headends, or at cell-tower sites. Their value depends on being close, in network terms, to the users or devices they serve.
Connectivity. A useful edge site connects to the local access networks, such as broadband, mobile and enterprise fiber providers, so traffic can reach users without first traveling to a distant hub. Some host a carrier’s point of presence (PoP) or a regional internet exchange point (IXP) where networks exchange traffic locally. The number of providers available varies widely from site to site.
Workloads. Common uses include content delivery network (CDN) caches, cloud and telecom providers’ local infrastructure, latency-sensitive applications and aggregation of data from IoT devices and branch sites. These are examples of edge computing running in a shared facility.
Design. Smaller sites may use prefabricated or modular builds and may be run with few or no on-site staff, relying on remote monitoring and dispatched technicians. Power and cooling redundancy depends on the operator; some edge sites match core facilities, others do not.
For help deciding where edge infrastructure makes sense, see our Edge Compute solution page.
When it matters for buyers
- Serving users in secondary markets. A local facility can improve application performance compared with a distant hub.
- Latency-sensitive or data-heavy workloads. Real-time processing, video analytics and similar workloads may benefit from nearby compute.
- Data residency. A regional facility can help keep data within a state, province or country, though residency depends on the full data path and provider practices, not location alone.
- Choosing between on-premises edge and a facility. An edge data center offers professional power, cooling and security without building them at your own sites.
Questions to ask vendors
- Where exactly is the facility, and what latency can we expect to the users or sites we need to reach?
- Which network providers are available in the building, and is there a local internet exchange or cloud on-ramp?
- What power, cooling and redundancy design does this specific site have, and does it hold any tier certification?
- Is there on-site staff, and how quickly can remote hands or a technician respond?
- What space and power increments are available, and how does pricing compare with your core facilities?
How it differs from edge computing
Edge computing is the model of processing data near where it is produced or used, wherever that is: in a store, on a factory floor, in a vehicle or in a nearby facility. An edge data center is a specific place where edge computing can run: a facility near users that provides shared space, power, cooling and connectivity. You can do edge computing without an edge data center, using equipment at your own sites, and an edge data center can host workloads that are not edge computing in the strict sense.