Enterprises across the world are rapidly digitizing their operations.
Increasingly, the digital strategies they are adopting include the transition of business tools, applications and processes to a ‘multicloud’ environment: essentially a hybrid combination of applications and data hosted in one or more public, clouds alongside the company’s own private data centres
With this shift, the traffic patterns and security postures of enterprise networks have been fundamentally altered. Multicloud has changed the world, and as much as the transition is impacting the inside of the modern data centre, it is also crucially reshaping the networks between them.
It thus comes as no surprise that enterprises for whom cloud and multicloud are a meaningful part of the architectural agenda will realize the need to expand their architectural ambitions beyond the clouds all the way across the wide-area network (WAN).
Change, however, is never easy. Organizations are indeed looking to make that transition, but often require the impetus for them to fully commit and make the jump. They are looking for solutions that would not just help them simplify their network, but also provide them with the flexibility and agility to evolve alongside constantly changing needs and requirements.
With this changing business landscape, it is almost inevitable that SD-WAN will eventually find adoption across the broader market. The need of the hour is to thus provide that pathway towards a secure and automated multicloud, in which organizations of all sizes can efficiently scale – from the smallest of organizations to the largest, most demanding and mission-critical enterprise networks of today.
If we looked at the networks of pretty much any enterprise customer today, the vast majority of WAN setups would still be on legacy architecture. While serviceable, the technical limitations are often significantly. Eight key ones among them include:
- The connectivity between the spoke (branch) and hub (headquarters or datacenter) is often on lease, while MPLS circuits are given by service providers.
- The circuits are always statically configured with one link as active and second link as backup. The backup link will be active only if the active link fails.
- The user in any of the remote location does not have the quality of experience from an application perspective, the bandwidth is increased or more links are added.
- All users from any of the remote locations can access the internet only through the Hub or DC.
- Management is very static in nature with manual provisioning for any move, additions or changes.
- The network administrator does not have any visibility of how the circuits are performing, or how the applications are treated over these circuits. Because of this, security measure also needs to be considered.
- Deployment of branches takes longer time and the complexity increases with the number of branches.
- For remote branches where the leased or MPLS circuits from service providers are not available, then connectivity of those remote branches will be not possible. An alternative using radio link to the nearest POP or VSAT connectivity will need to be used.
When factored in together, besides the above there will also always be significant increases in CapEx and OpEx spending, due to the growing number of connectivity leases, MPLS circuits and management overheads required, all of which will need a greater number of skilled engineers to manage the network.
With network modernization in mind, organizations are increasingly mapping out their transitions towards the multicloud. However, in the meanwhile this often still leaves the corporate data centres and hub sites having to deal with rapidly increasing bottlenecks. In addition, their existing static and manually provisioned circuits often just can no longer keep pace with the dynamic nature of multicloud traffic flows.
To ease the transition, organizations should consider a ‘network for the future’ approach, and a step-by-step process.
Consider utilizing solutions that ease the burden of programmable network devices, which you can modify remotely and through dynamic best-path routing. Both of which improve cost, agility and performance, and also respond to real-time network condition.
Do also consider viewing disparate pools of resources into a single, cohesive entity that reaches from the cloud to cloud on-ramp, from the data center to the branch – and yes, that includes the WAN. This requires stretching of the operational domain over what has traditionally been siloed networks.
In the next part of this article series, I will be detailing the step-by-step approach that organisations must focus on to ensure a successful transition.
By Rajeshkumar S, Head of Systems Engineering Enterprise & Govt. Juniper Networks India