Over the last couple of years the demand for connectivity has changed and entered a new phase. Traditional networks were centered around offices with on premise IT-infrastructures complemented by a back-up solution in an external data center. The most common connectivity solutions were dark fiber, WDM, ethernet, MPLS and IP VPN.
Cloud has changed all that. Since its entrance into the connectivity ecosystem it has quickly become clear that traditional connectivity solutions often are insufficient to fully benefit from cloud solutions. Many organizations have experienced poor cloud connectivity and look to their trusted local IT-partners to offer them a reliable, secure and cost-effective solution. How can managed service providers fulfill these new connectivity demands in an effective way?
Increased importance of connectivity
As more and more workloads are moved to the cloud, connectivity to office locations of enterprise customers has become increasingly important in the totality of the IT-infrastructure. Unstable networks, bandwidth restrictions or packet loss all result in poorly functioning cloud applications.
While the traditional networks were all under control in local facilities and the data center of an enterprise, connectivity to the cloud is dependent on the location of the cloud provider, the way cloud providers provide access to their cloud and the way local network providers provide access to the cloud provider.
The limitations of cloud access via public internet
Microsoft Azure, Amazon Web Services and Google have each standardized their cloud access in their own way. The preferred way to connect customers for these hyperscale public cloud providers is the public internet. It is easy to implement, highly standardized and available everywhere. But it also has some serious drawbacks.
As more workloads and business critical applications are moved to the cloud, performance issues are increasingly likely if the internet is the primary way to connect to the cloud. Packet loss and jitter are serious risks, peak traffic may result in throughput issues and there is little to no control over routing of traffic. Also security risks are raised as data traffic that has been private in traditional situations now is supposed to be transferred via internet which is inherently unsafe without the right precautions.
Most managed service providers do not consider connectivity as a pivotal part of their services. As vendors of cloud services however, their customers turn to them when application performance or security issues arise. Connectivity plays a critical factor and that’s why we believe there is an opportunity for MSP’s to help their customers get the right connectivity for their cloud needs.
Adapt, avoid or all-in?
There are, of course, ways to fix these issues. One of them is rearchitecting networks. Steering away from a traditional focus on office location and data center connectivity may help to handle workloads going into and coming out of the cloud. But this would be a costly, time-consuming operation. A network has to become more flexible and probably should be partially or fully automated to be able to handle agile workloads. In many cases, this option simply isn’t viable.
Another option would be to stay away from cloud. However silly this may seem, recent data on cloud adoption provided by industry association Dutch Hosting Provider Association (DHPA) suggests this strategy is actively pursued by some organizations. The reason? A (perceived) lack of security and control, which also has to do with connectivity.
Last but not least, from an end user perspective, it might seem a good idea to put all eggs in one basket and exclusively do business with one CSP. However, this will put a serious strain on an organization’s’ agility and capability to adapt, not to mention there would be a real risk of vendor lock-in.
Managed service providers should always consider the customers’ needs and keep an open mind to any solution, even if this would require choices that are unorthodox. In most cases, however, cloud will probably be part of this solution because customers need the flexibility the cloud offers. Because of this flexibility, customers will also have multiple and quickly changing ways they use cloud services, which also requires agile cloud connectivity. Flexibility of contracts, bandwidths, connected locations and access to different cloud providers thus all become more important to organizations.
An end-to-end service
As we have seen, cloud connectivity can be quite a hassle. We therefore use a standardized DCspine Cloud Connect for end-to-end connectivity based on a L2 private ethernet connection. In this way, we are able to offer a single SLA and take care of every hop and hurdle that is encountered along the way. And best off all: we are able to deliver these connections on demand from any data center we have a point-of-presence in making sure customers can fully benefit from the cloud whenever they want.