As a new technology bursts onto the scene, it is common to see familiar terms being used — or in some cases misused —along with erroneous associations to explain its value and benefits to the marketplace. For example, a rotary dial telephone opened up the world of personal communications, but there is no real comparison between the inner workings of those earlier devices and the digital smartphones that have revolutionized the way we communicate and interact with the world today. Yet, we call them both “phones.”
“Cloud peering,” a misnomer that is often mistakenly equated to “network peering,” is a very timely example of how two similar terms have been used to describe things that solve different problems in fundamentally different ways.
In network peering, there is an equal relationship between partners. For example, two large Internet service providers (ISPs) share networks to move their traffic to their customers over longer distances. In most cases, these equal parties agree to use “settlement-free peering” to exchange equivalent amounts of traffic between themselves or agree to a compensation model that is proportional to the balance of traffic.
However, this type of peering is not representative of what is happening when an enterprise establishes a private, direct and secure interconnection between itself and a public cloud giant, such as Microsoft Azure or Amazon Web Services (AWS). That is not really a “peer” relationship. In fact, it is a very different commercial model than two ISPs or networking partners that are of a similar size and in the same business.
In contrast to a peering model, the relationship between the enterprise customer and the cloud provider is not equal and requires a different interconnection model. The cloud provider wants to deliver an easy, faster and more secure way for customers to access their services. The old interconnection model of network peering is really suboptimal for this type of relationship because most of the data traffic is traveling between the enterprise and the cloud provider “indirectly” over the public Internet. The Internet cannot provide the security, throughput and low latency required to maintain a high-level, quality relationship between the cloud provider and enterprise customer. This can only be achieved through a direct, private connectivity model.
To illustrate the differences in interconnection performance when you connect to cloud services directly versus over the public Internet, let’s look at the test results Oracle and Equinix achieved when they used Oracle Fast Connect and the Equinix Cloud Exchange. Oracle and Equinix recently announced that they are enabling enterprises to receive dedicated, direct and secure access to Oracle Public Platform and Infrastructure Services via the Equinix Cloud Exchange.
In a database backup test scenario to the Oracle Public Cloud, a 1 terabyte database was replicated over the Internet via a 1 Gbps connection, and then compared to the time it took to back up the same database via the Cloud Exchange — using first a 1 Gbps connection and then a 10 Gbps connection. Here are the test results:
In addition, tests were run comparing access to the Oracle Cloud Compute Cloud Service in which the Equinix Cloud Exchange delivered up to a 60-fold improvement in backup time performance as compared to going over the Internet, with 27% less latency and 40% fewer network hops. These tests demonstrate how direct interconnection between cloud service providers and enterprises can be optimized seamlessly across an enterprise and public clouds.
Much like the giant leap in personal communications between rotary telephones and mobile smartphones, at Equinix, we are enabling a dramatic shift in public cloud interconnectivity. By providing both private cross connects and the Equinix Cloud Exchange, we are moving the industry toward a private cloud connectivity model that enables high-performance, secure hybrid and multi-clouds.
Learn more about high-performance multi-cloud interconnectivity via the Equinix Cloud Exchange.