gluke - Fotolia

Harnessing hybrid deployment

Computer Weekly explores how the success of hybrid deployment hinges on the network

Businesses increasingly depend on cloud to help deliver differentiated services to customers.

As a consequence, rampant adoption of new infrastructure platforms has changed the composition of the enterprise network, altering provider solutions, consumption models and performance-driven architectures.

Gone are the days when telecom firms were your only choice for network services. Cloud giants, colocation providers and wireless services are among the new choices that will augment (but not replace) the telecoms titans.

Communications services are already moving from fixed pricing and long-term contracts with slow changes to models that flex quickly with equally agile pricing – similar to cloud computing.

The connected enterprise

Connectivity is the central nervous system of the business and cloud adoption has changed the network’s composition. Intense cloud adoption continues to transform service design and delivery, and places additional strain on networking infrastructure.

Some 83% of enterprise network and telecoms decision makers at enterprises have already responded to local performance challenges with network hardware and software upgrades as part of private cloud initiatives, but organisations must also address challenges introduced by integrating various external resources.

Compounded with increasing bandwidth requirements, hybrid cloud pursuits have exaggerated the limitations of the network beyond the datacentre. Integrations among an increasingly distributed portfolio of applications, services and data require new optimisation strategies at both the component and aggregate layers. 

Distinct products, suppliers and strategies have traditionally served datacentre and branch office network infrastructure. However, organisations can’t build cohesive solutions atop disparate networking segments.

To address this, a single virtualised business-wide fabric has emerged to support the new digital business network.

As customers increase their use of cloud platforms and expand to multiple regions, cloud providers will be responsible for a larger portion of their networks. And as cloud providers continue to increase the number of direct connect points of presence (PoPs) and lay proprietary connections between their facilities, customers have more options to leverage cloud network infrastructure and controls in and between these datacentres.

Network disrupters

Cloud adoption has expanded the heterogeneity of transport technologies in the enterprise network. Software as a service (SaaS) adoption accelerated broadband use at the branch office, and cloud direct connections have increased to both proprietary datacentres and branch office wide-area networks (WANs).

Software-defined WAN (SD-WAN) solutions have emerged to abstract this complexity, in turn providing a software-defined overlay to manage multiple connection types as well as automate traffic between links per defined application policies.

While many carriers sell this kind of solution via a white-label model, the intrinsic capabilities of these platforms enable customers to take back control of the WAN, and reduce dependency on multiprotocol label switching (MPLS) and managed solutions.

With growing pressure to design and deliver dependable services that provide compelling digital experiences, businesses have already begun to shift network design and supplier relationships to accommodate these demands.

Many expected cloud service adoption to simplify networking, but it has only accelerated network evolution in positive (but complex) ways.

Cloud-ready networking

Cloud has reset expectations of infrastructure flexibility and puts strain on inflexible network components and services. To achieve true cloud benefits, the network must be adaptable to real-time traffic patterns and scale capacity allocations to meet changing demands.

While traditional connectivity offerings are rigid, many actively follow Forrester’s virtual network infrastructure tenets to increase flexibility, simplify management and extend granular controls and application programming interfaces (APIs) to customers.

Incumbent telcos aren’t the only providers of connectivity services. Increasingly, customers are purchasing connectivity through myriad providers: systems integrators, resellers or other types of suppliers. 

Colocation and managed infrastructure services suppliers often bundle their own connectivity capabilities, and some have gone even further to offer aggregated solutions, managing routing across the best connection at any given time. 

While carriers still sit below many of these services, they’re further removed from the customer and more easily swapped out in favour of alternatives.

Despite constant pressure to consolidate and reduce sprawl in the infrastructure environment, multi-cloud adoption continues to distribute infrastructure resources across a wider variety of providers and locations. As resources become more distributed, increasing traffic between a variety of internal and external components will strain hub and spoke models (star topologies) designed for legacy traffic patterns.

As a result, hub and spoke WAN design is eroding, giving way to a more richly interconnected mesh-fabric WAN to support hybrid infrastructure environments. 

Emerging solutions are going beyond reselling carrier services, with more of them offering abstracted, aggregated or proprietary alternatives for connectivity.

As the major infrastructure-as-a-service (IaaS) providers continue to expand, they are investing in both datacentre capacity and network connectivity between their facilities.

Because they provide globally distributed platforms, intersite connectivity is a key component of their services, so owning and managing the network layer allows for full system engineering and optimisation.

Content delivery networks (CDNs) optimise routes and caches over an aggregated network. Colocation providers bundle carrier services and offer blended IP products.

With cloud services inflating expectations of speed and agility, many telecommunications providers have started to invest heavily in virtual network infrastructure (VNI).

However, due to the scope of requirements to implement VNI and recognising limited internal resources, many providers have put aside legacy feuds and looked to open source communities to collaborate with their peers, including the Open Compute Project, OpenDaylight and OpenStack. These serve as mechanisms to jointly innovate and move the industry further and faster.

In addition, APIs open telcos up to new opportunities for digital integration with enterprise customers and ecosystem partners.

The changing role of networking

Infrastructure and operations teams will always be responsible for service quality and availability, regardless of who owns and operates the service.

As organisations adopt multiple cloud services across their software and infrastructure, they’re not only strategically replacing in-house capabilities but also increasing the number of providers nested in their service delivery models.

More suppliers translates to increasingly distributed responsibilities, and multiple deployment types indicate multiple hand-off points.

While service-level agreements will state performance requirements, IT professionals must monitor this heterogeneous supply chain to hold suppliers accountable to their requirements.

This article is based on the Forrester report, “Adapt your network strategy to thrive in a shifting ecosystem”. Sophia Vargas is an analyst at Forrester.

Read more on Software-defined networking (SDN)