Software-Defined Networking & Interconnection

Next-generation applications require a next-generation networks

Vapor IO’s Kinetic Edge and Kinetic Grid networks must support microsecond latencies, automated end-to-end provisioning, and API-level access to telemetry data and control. Data must travel over predictable, high-speed, resilient routes. Network-to-network cross-connects must happen in real-time. Applications and services just must be able to make real time changes to the network in response to outside events.

Home / Technology / Networking and Interconnection

Hyperscale-Grade Networking

Bringing the core of the internet to the edge

Today’s internet was built from the core out, not the edge in. The next generation of applications that operate and process data at the edge will require levels of network performance that have traditionally been relegated to large tier 1 and hyperscale providers moving data between major internet exchange points. Vapor IO employs the same hyperscale-grade technologies used in constructing the internet’s core and we bring those network capabilities to the edge. Every Kinetic Edge market provides a middle-mile backbone network that offers sub 1ms latencies between facilities.

Software-Defined Everything

The network of the future operates as software, which is why Vapor IO technology lets customers set up, configure, and manage private backbone networks through custom-built APIs

Edge-to-Anywhere Routes

Vapor IO’s platform tightly-integrates with the first and last mile networks. In-market (local) is routed efficiently via the Kinetic Edge; data destined for other markets can pass through internet exchange points or travel on private backbones.

Latency-Aware Topologies

Next-generation applications require deterministic network routes with extremely low-latencies. Vapor IO builds its networks for the lowest-possible latencies within and between markets.

Hyperscale Grade Backbones

Vapor IO builds metro and national backbones by lighting dark fiber with PAM4 optics, employing high-end core routers, and configuring software-defined private networks using the latest VxLAN technologies.

Network-Centric

We embed our data centers at strategic locations within our high-capacity metropolitan backbone network. Our data centers become part of the network and support local transit and interconnection in ideal locations with minimal hops and predictable performance

Optimal Locations

We carefully analyze existing and anticipated fiber networks, the locations of last nuke cellular and cable facilities, along available land and we plan an initial configuration to optimize for latency and resilience. But we also plan years into the future so that we can easily support extensions and densification of our Kinetic Edge architecture in each market

Kinetic Edge Architecture

Vapor IO’s Kinetic Edge architecture is the unique combination of high-speed software-defined networking, edge colocation and edge exchange — all pre-integrated into a single package from a single supplier—lowering the cost and complexity of edge computing.

Latency-Sensitive Topology

In any given market, the target latency between adjacent facilities is less than 1 ms, making geo-redundant failover a realistic option. In markets where our customers are building 5G systems, our target latency to remote radio heads and adjacent facilities is less than 100 µs.

Kinetic Edge Network

Software-Defined Metro Networks

Each Kinetic Edge market has a custom-built Kinetic Edge network that behaves like a hyperscale backbone. These Kinetic Edge networks connect the in-market Kinetic Edge facilities to each other and to other important destinations, including prominent last mile networks, regional colo facilities, and nearby internet exchange points. The Kinetic Edge network can be extended to on-premises facilities, allowing enterprises to offload on-premises workloads into near-premises facilities without sacrificing latency or privacy.

The Kinetic Edge

Like a City-Scale Data Center

The Kinetic Edge architecture is built around the idea of scaling a single data center to the size of a city. Imagine a typical data center with rows of racks filled with servers. Each of these racks has a Top of Rack (TOR) switch and connects to other racks in the data center over a high-speed packet network. Now further imagine splitting those rows apart and placing each row in its own building, then moving the buildings apart—but keeping them close enough to each other that the latency is less than 1ms. The servers in different rows, though they are now separated from other rows by miles of fiber optical cable, would still behave as if all the rows and all of the servers were still in the same building. By maintaining low latencies and using a campus-style network, a centralized data center’s worth of servers can be geo-distributed into smaller facilities spread throughout the metro—this is the Kinetic Edge

Kinetic Edge Exchange

Machine-Speed Interconnection

The Kinetic Edge Exchange (KEX) cross-connects any two networks via software and at machine speed. This virtualized cross-connecting fabric allows Vapor IO customers to make real-time business decisions about how they virtually interconnect and exchange traffic. Any two networks can be cross-connected or disconnected, on demand, through network APIs. The KEX unleashes interconnection at the local level and democraties network-to-network interconnects.

A Backbone that Spans the Nation

National Connectivity with Zayo Dark Fiber

Vapor IO has partnered with Zayo, one of the world’s largest providers of long-haul fiber, to build a nationwide network that connects Kinetic Edge markets to each other via Vapor IO’s technology. This allows Vapor IO customers to create flat networks that span markets and to achieve the lowest-latency and most cost-effective routes for data that needs to leave a market.

Low-Latency Applications

Built for 5G Networks

One of the biggest shifts between 4G and5G networks is the upgrade to virtualized network functions (VNFs). No longer locked into physical appliances, VNFs allow telco system operators to run their network on readily-available cloud servers. The 5G Radio Access Network (RAN) requires very precise latency budgets to operate correctly. Vapor IO has designed its Kinetic Edge networks to support the 5G RAN 7.2 split, which requires 100µs or less latency between the virtualized baseband unit and the tower that carries the radio head. This allows the Kinetic Edge to support virtualized 5G network functions in low-latency environments alongside application servers that provide low-latency services that utilize the 5G networks, minimizing any route inefficiencies for exchange data between the application and the last mile network

As Seen In

The State of the Edge

Download the state of the edge report today

The State of the Edge report is the go-to resource to understand edge computing in the industry today, and is created by experts in the field. Covering edge computing terminology, architecture, technologies, applications and futures, don’t miss your chance to read it now.

Consult with our sales team

Resources

What's New

Blog Sep 21, 2022

Next-generation applications require a next-generation networks