The Benefits of Data Center Interconnect for the Cloud

This is the final in a series on data centers and data centers and data center interconnect (DCI), based on interviews and white papers provided by Gary Holland, IP Routing and Transport (IPRT) Marketing with Alcatel-Lucent (News - Alert) (soon to be part of the new Nokia) and his team. In the first post, Holland provided his top four trends for the coming year.  The second posting dove into why, when it comes to having the right cloud and DCI solution for your organization, getting the balance right along with having solutions optimized for your unique situation is crucial.  In this final installment, focus is on the benefits to large enterprises, as detailed in the Alcatel-Lucent white paper, Cloud Interconnect for the Enterprise: Data center interconnect for the cloud era, of moving to the cloud with a sound DCI strategy in place.

Putting the DCI strategy for cloud in place

In the previous posting the criticality of having the right DCI strategy was discussed, based on the undeniable history of the communications industry that all boats really do rise when the tide comes in. In short, increasingly higher and higher up the stack, in what will be a data center-centric world, interconnect is not a nicety but rather is a necessity. It was also noted that unfortunately not all data centers solutions are alike, not all clouds are alike, and thus, finding the DCI strategy that works best for any organization can as Alcatel-Lucent says, “can be a sometimes-daunting and complex task, even for a well-resourced large enterprise.”

A primary culprit leading to this complexity is the sheer speed in the growth of cloud IT where the volume of data exchanged between cloud data centers is exploding exponentially. In particular this has put significant stress between an enterprise’s primary and backup data centers located in the same metropolitan areas. This is not to mention those located more remotely to mitigate risks during major disasters and/or to be closer to important customers and communications hubs.

Just looking at metro network traffic growth, for example, Alcatel-Lucent Bell Labs (News - Alert) predicts that traffic between enterprise data centers in metro networks will increase by 430 percent from 2015 to 2020, due almost entirely to the adoption of cloud IT. Business continuity is one reason but so is the fact that most applications are now being developed to reside in the data center/cloud.

What this has translated into is a need for speed and the need to seamlessly interconnect. In response many large enterprises are using optical wave division multiplexing (WDM) solutions. These provide the scalable bandwidth needed along with the price performance and low-latency required for optimal operations.

As ALU points out, today’s large enterprise DCI applications require bandwidth what not that long ago would have been considered almost futuristic and is now almost obsolete, between 1 Gb/s and 10 Gb/s. However, they note that the following factors can place an increasing strain on enterprise DCI networks, requiring bandwidth up to 100 Gb/s and beyond:

• Convergence (News - Alert) within the data center: The eventual migration of data center network fabrics to 40 GE and 100 GE in the core will cause traffic to increase within and between data centers.
• Adoption of cloud IT: As data center server and storage virtualization increases a request in one data center will more than likely trigger multiple data exchanges between servers in one data center, as well as servers in different data centers.
• Rapid advances in storage technology: Because of the rapid pace of storage innovation, the case for more cloud-based storage has become compelling.
• Continuous data availability and mobility: As virtual compute and storage resources are separated and distributed across physical devices and locations, DCI solutions may require more bandwidth to ensure business-critical applications remain up and running at all times.
• Dynamic allocation of resources: As IT is virtualized within and across clouds, dynamic allocation of server, storage and network resources across cloud boundaries becomes essential for resource sharing, application workload balancing and greater operational efficiency.

The benefits of cloud interconnect

What should resonate with readers is the observation in the white paper that: “Enterprises are moving toward building a complete digital infrastructure – one that includes using a policy-based approach to orchestrating the use of compute, storage and applications. Cloud interconnect solutions have a role to play in helping enterprises dynamically match and execute workloads at the best location, based on cost, performance, security and other key business requirements.”

That is a good summation of why having a well-defined DCI strategy, and developing and implementing with some sense of urgency I would add, has become important for large enterprises and their service providers. The explosion in traffic like Pandora’s Box (News - Alert) cannot have the lid put back on.  More importantly, the benefits to large enterprises when using dedicated cloud interconnect to link their private cloud to virtual private and hybrid cloud services provided by CSPs, CNPs and ICPs include:

• Security: A dedicated cloud interconnect link is more secure than the public Internet. When combined with existing private WAN backbone and managed services, branch offices and remote locations can use cloud resources.
• Cost: A private cloud interconnect link can reduce costs. Traffic does not have to be routed over an ISP’s connection to the Internet. Instead, it’s transported directly to the cloud provider.
• Performance: Bandwidth (News - Alert), latency, response time, quality of service and reliability are more consistent with dedicated cloud interconnect.
• Flexibility: Access to a variety of cloud services can be implemented over the same dedicated cloud interconnect link. This means different workloads can be allocated to resources that have the appropriate price/performance profile.

Finally, and it should be noted that the sense of urgency I described in having the right DCI strategy to go forward with means have more than a road map.  It means have the technology, that can meet the attributes for security, cost, performance and flexibility (I would add manageability, visibility and control to the list as well) in this list above.  And, not surprisingly ALU has a large portfolio of solutions for large enterprises that enable them to have the DCI solutions for cloud interconnect that best fit their needs.

That said, it is important to return to what in many ways has become obvious to large enterprise IT departments as they grapple with network transformations that best align with rapidly changing business imperatives. Realities are that in a software/data center-center world failure is not an option. Hence, putting decisions about DCI cannot and should not be considered something to delay.