Enterprises are becoming more reliant on networking, whether they realize it or not, as a byproduct of their reliance on public clouds and SaaS providers. This is especially true in sectors such as financial trading – consider, for example, the firms that rely on high-speed connectivity to hub cities such as Chicago and New York. Low latency is crucial and, beyond that, the need for security and reliability is heightened compared with most enterprise work. This is not a job that can be handed to the public internet. The situation is best handled by private connectivity, creating a more secure, reliable network where low latency can be assured.
This would be a relatively simple proposition if the enterprise had to connect to only one destination, such as a trading desk. But enterprises in all sectors are pursuing hybrid IT, combining public clouds with enterprise-controlled facilities – whether in colocation (including private clouds) or truly on the enterprise premises. In a recent 451 Research survey, 46% of enterprises said they have a hybrid IT architecture in place, and another 16% are in the process of implementing one.¹ Add to that the fact that most enterprises are already using more than one public cloud and are increasingly dependent on cloud-delivered services (SaaS providers, in other words), and you get a complex and growing connectivity profile to be maintained.
It's likely that this profile has been built up ad hoc as different cloud services were added to the enterprise portfolio. Moreover, it's likely to fluctuate over time, as the enterprise makes changes to its cloud usage.
One way to wrangle this networking is through colocation. A leased data center in the right location becomes a crossroads for all manner of network providers, in addition to supplying on-ramps to the public cloud. From that data center, the enterprise can establish private connections to a multitude of venues, assuring itself of the performance and security that comes with this connectivity. The key is that the creation of these connections – the cross-connect process – must be automated and virtualized. This is where software-defined networking comes in: once a physical cross-connect is in place, an automated process lets the customer activate virtual cross-connects at will, without waiting the traditional weeks or months for a work order. In a data center context, we at 451 Research refer to this functionality as software-programmable interconnection.
This is how networking joins the ranks of cloud-native infrastructure, where resources can be automatically scaled and pooled, and it's how networking operations can keep up with the pace dictated by the financial world. For a financial firm looking to use connectivity to its advantage, software-programmable interconnection can provide a ready-made platform for fluid, controllable networking.