App varieties: The spice of life
Network analytics and today's changing app landscape bring increasing challenges for network managers and administrators.
Millions of new software applications have been developed to support evolving personal demands and business requirements and efficiencies. As of March 2017, Android users were able to choose between 2.8 million apps, while 2.2 million were available from the Apple App Store. Globally, thousands of new apps are added each day.
Today, network managers and administrators face unprecedented challenges linked to the increasing varieties of applications available, with different and often specialised requirements for delivery, encompassing traffic priority, latency and security.
Because of app proliferation, the level of performance required from the corporate network needs to be significantly revised - particularly in terms of quality-of-service bandwidth - in order for it to adequately fulfil the demands of the business it serves.
Modern app world
Against this backdrop, moves away from traditional applications - which have to be installed and maintained - to virtualised, next-generation apps complemented by private and public cloud-based delivery models (such as SalesForce.com, Google Apps and others) are making dramatic and lasting changes to the app landscape.
In addition, the rise of concepts such as "mobility" and "agility" across all aspects of the enterprise has resulted in users demanding immediate, on-demand access to all of their business and social media-related apps, no matter what device is in use.
The increasing implementation of virtualised and cloud-based resources that make it possible to distribute applications across a comprehensive range of devices has led to a steady increase in network complexity. This has also been driven by technologies related to the Internet of things (IOT) and the mounting numbers of networks adopting wireless communications.
Changes in the fabric of the corporate network to accommodate the complexities these advances bring are raising issues associated with effective network management. This is exacerbated by the lack of any permanent and exclusive relationship between modern networks' physical resources and the software that runs on them.
Traditional enterprise networks have played a limited role when it comes to optimising application delivery, because they transport "raw information" from one place to another without any regard for an application's individual requirements.
Thousands of new apps are added each day.
Previously, network designers created networks with enough raw bandwidth to meet instances of peak application delivery in an approach known as "over-provisioning".
While over-provisioning strategies have proven successful, they represent a costly and increasingly outmoded architectural model, particularly in the budget-conscious, modern-era.
To effectively understand virtualised application behaviour, network managers need to gain higher levels of visibility and control of their networks in order to apply policies effectively aligned with evolving business goals.
The answer lies in the emerging and constantly refined technologies underpinning network analytics.
Network analytics, in its simplest definition, involves the analysis of network data and statistics to identify trends and patterns. It enables companies to optimise network infrastructures and operations, thus reducing costs in terms of both capital and operational expenditure.
It is also able to capture and analyse context-based application traffic to deliver meaningful intelligence about applications and their users together with data relating to users' personal computing devices and geographic locations.
For network managers to fully understand and influence the behaviour of applications, application monitoring via network analytics is key. Importantly, it also enables managers to gain the required levels of visibility into their networks and exercise the control they require over their infrastructures.
In this light, modern analytics products are now available to help managers make optimal business- and network-relevant decisions in an efficient manner and provide sound platforms from which to address the challenges and opportunities introduced by billions of connected IOT devices and associated digital services.
Significantly, network analytics now allows network managers to create security, policy or behavioural alerts tied to specific values extracted from the results of a saved search.
From a technical perspective, network analytics solutions facilitate custom aggregations, multiple chart formatting options, real-time dashboards and historical views through trend reports, while taking into account the variations in terms of source and format of the data to be analysed.
Looking ahead, network analytics will provide the impetus behind the future development and deployment of new statistical analysis techniques and methodologies.
These will be employed to identify dependencies between commercial and technical variables through the use of ever-refined predictive techniques such as simple/multiple regression and projection. Proactive, reactive, behavioural and contextual conclusions, unparalleled in scale and performance, will be provided.
Paul Stuttard is a director of specialist distributor Duxbury Networking. Currently Cape-based, he has been with the company for 29 years and has extensive experience in the IT industry, particularly within the value-added distribution arena. His focus is on the formulation of future-oriented network optimisation strategies and business development objectives in collaboration with resellers and end-users in Southern Africa.