Many enterprises today are embracing a hybrid IT strategy, adopting public clouds while retaining private clouds that may or may not be hosted on-premises. Where servers and storage once resided centrally in the enterprise data center, those resources are now distributed outward towards the cloud. As workloads migrate to new platforms, proprietary applications might move to the public cloud and legacy applications remain on-premises, resulting in a fragmented infrastructure.
The increasing complexity of the cloud can often prove to be an obstacle to fully unlocking its benefits. The migration of business-critical applications can lead to problems that require enterprises to re-architect their network. In doing so, it’s likely they’ll encounter several challenges – cost and network impact among them – as they work their way through the waves of Enterprise transformation.
As cloud complexity increases, many enterprises have made incremental and nonstrategic investments in their networks. These investments have often resulted in complex and inefficient infrastructures, characterised by high fixed costs, limited scalability, and an inability to effectively support IP-based business applications.
For example, MPLS used in conjunction with SD-WAN has long been the popular architecture of choice – particularly in the enterprise – but even this approach has its issues. MPLS lines are expensive when dealing with high volumes of data transfer, and SD-WAN often sends more traffic over the internet to reduce the strain on MPLS, compounding the lack of network control, latency, and lost packet issues.
Scaling the required network resources can prove challenging too and attempting to do so can lead to valuable focus being diverted away from managing an enterprise’s core business. As enterprises make their way ever closer to the third wave of transformation, and the additional demands this places on their networks, greater support is needed in re-architecting those networks to make them future-proof for managing the rising volume of data and interconnecting workloads.