Enterprise Networking – Then and Now

7 Enterprise Networking Trends for 2018 | Network Computing
https://www.networkcomputing.com/sites/default/files/styles/flexslider_full/public/1-intro_96_0.jpg?itok=wxG9CE81

For decades, networks has played critical role in connecting users to applications whether its the enterprise networks or the giant Internet as we know it. While the network security , resiliency and performance was not that critical for the Internet but for the enterprise networks these aspects translated to lost revenue for the company. We as network engineers can brag all night as to how difficult is the traditional networking construct from functionality to manageability. So I am not getting into this. But a case in the point: Back in 2010 it took me 2 days to troubleshoot a QoS (Quality of Service) Issue for a bank in Malaysia.

Over the past decade we saw the shift in parts of enterprise application workloads moving from on-premise data center to the multiple clouds and at the same time user mobility went beyond enterprise boundary. In the current state, enterprise users connect from anywhere (home, coffee shop, on road, on-premise…) and at the same time there is a proliferation of IoT devices needing new connectivity models to the cloud. This has fueled unprecedented innovation In the enterprise networking space with the sole purpose of securely connecting any device, any location to any data center or cloud. Essentially we are talking about any to any connectivity model. As client-server traffic patterns has changed due to this shift in the application space, the networks had to adopt to support new connectivity models with the same level of performance and security as traditional networks.

Along with the new connectivity model what has also changed is the dis-integration of various network functions and addition of new ones for centralized policy creation, granular control, manageability and orchestration. On a broad level it is termed as Software Defined Networking (SDN) which in simple terms is the separation of control plane (brains of the network node) and data plane (just forwards the traffic) layered with additional management and orchestration layers.

Whether its the WAN, campus or data center networks, all these network types are software defined today allowing administrations to define centralized policies to the network nodes from a single location. For example, Software Defined Wide Area Network (SDWAN) solution designed for enterprise branch connectivity to multiple clouds through MPLS, broadband and LTE , has four different planes of operation: Data plane, control plane, management plane and orchestration plane. What’s more interesting is that except for the data plane the remaining planes reside in the cloud acting as brains of the SDWAN solution. A rough analogy here is the central nervous system (brain and spinal cord) controlling various bodily functions (this can be compared to data plane operations).

Software defined networking is a dire necessity for companies engaged in digital transformation. While the software defined network has enabled better control through centralized policies, Increased automation, user and device mobility and better performance, there is next level of Innovation happening in this space.

This next level of Innovation has to do with the multi-cloud networking. Companies who currently has application workloads across different clouds (AWS, Azure, Oracle cloud…) are increasingly dealing with complexity of operating various network constructs as each of them are unique and has their own nuances. At the same time, the current co-location model of connecting enterprise to public cloud through exchanges like Equinix is presenting some unique challenges. The solution to these problems is called Network Services Exchange, which is an abstraction layer from all the different cloud including data centers. Without worrying on the underlying networking constructs of multiple cloud providers this exchange can be used to create and manage public cloud network with a drag and drop like feature on a canvas. Please refer to this article from Williams Collins for in-depth information.

Leave a Reply

Scroll to top
%d bloggers like this: