Sign up for our newsletter and get the latest big data news and analysis.

For Service Providers, Big (network) Data is the New “Oil”

In this special guest feature, Alex Pavlovic, Director, Product Marketing, Nokia Deepfield, discusses “Data is the new oil,” a phrase coined by The Economist, that has become the mantra for describing the platform-based economy driven by hyperscale players such as Google, Facebook, Amazon, Microsoft and Netflix. Alex’s telecommunications career of more than 25 years spans many environments: academia, regulatory, consulting, and Tier-1 hardware and software telecom vendors. Alex started his career working on academic and research projects at the University of Belgrade, followed by harmonization of ITU-related standards and national regulatory activities. Currently, Alex is a Director, Product Marketing in Nokia, focused on the Nokia Deepfield portfolio of applications for network intelligence, analytics and DDoS security. Alex obtained his B.Sc. and M.Sc. degrees from the Faculty of Electrical Engineering (ETF) at the University of Belgrade, Serbia.

“Data is the new oil,” a phrase coined by The Economist, has become the mantra for describing the platform-based economy driven by hyperscale players such as Google, Facebook, Amazon, Microsoft and Netflix. These companies deliver valuable services to end users and consumers, but they also collect vast amounts of data and use it as a foundation for many of their business models.

To reach their global audience, these webscale giants need the ubiquitous connectivity provided by service providers and their networks. These networks generate and store vast amounts of data on networks, infrastructure and services. For example, the evolution to all-IP and exponential growth of traffic, devices and users have turned internet routers and other networking systems into massive generators of control plane information in the form of streaming telemetry.

For service providers, the challenge is to turn this data into valuable knowledge about their networks, services and subscribers. In many cases, service providers can’t use large volumes of data that may be available to them because they lack strategies or adequate tools for data mining.

However, big data-driven networking has already been recognized by industry and standards organizations such as ITU as the model architecture for future networks. This model enables service providers to use intelligence from big data to facilitate automation, streamline operations, improve quality of experience (QoE), strengthen security and increase business agility.

Know your network, but know the internet, too

Most of the content that service providers deliver to their customers is internet-based. Video streaming and gaming generate the lion’s share of network traffic, and traditional voice and messaging services are making way for internet-based equivalents such as WhatsApp and Viber. Service providers that want to be successful today must deliver a premium QoE for over-the-top (OTT) services, too. This demands precise knowledge about where and when to make network changes such as adding content caches or optimizing peering and transit.

To know how best to deliver internet services, service providers need solid insights into how internet traffic traverses their networks – from peering and transit interfaces, across the IP core and backbone, to network edge and access layers. They need a full perspective on the internet and its content-originating domains and content delivery networks (CDNs), along with the ability to identify and classify traffic flows. Traditional approaches based on network appliances can’t be easily or cost-effectively scaled for the cloud era. Approaches based on deep packet inspection (DPI) technology can be foiled by end-to-end traffic encryption.

Service providers need a new approach that combines detailed knowledge about internet services with full network and service insights. This approach must maintain a precise, dynamic “supply map” of internet services and applications (e.g., Deepfield Cloud Genome) and correlate it with big data from the service provider’s network in real time.

With an approach that fully correlates these two large data sets – using an open framework that makes it easy to add other relevant data sets – service providers can get multidimensional insights about how internet and on-net services are provisioned, delivered and consumed. Without this holistic perspective, service providers will not be able to derive benefits from vast volumes of network data that may be available to them.

ML, AI and big data-driven networking

Machine learning (ML) and artificial intelligence (AI) are instrumental for capitalizing and monetizing big network data in service provider networks. Their implementation will follow the evolutionary path shown in Figure 1, but it will only be effective if it applies equally to both internet-related and network-related realms.

Figure 1: Evolution from big data to augmented intelligence,
taken from “The Future X Network: A Bell Labs Perspective,” by Marcus K. Weldon

For the internet-related data, many ML algorithms are already in place, augmenting early and pre-ML heuristics. These algorithms cover:

  • IP classification
  • DNS matching
  • Traffic identification and classification into categories
  • DDoS security, where all kinds of behavioral and pattern-matching techniques are in use

At the same time, service providers need to apply ML techniques to network-related data to support diverse parts of their organization:

  • Network engineering and capacity planning
  • Service planning and marketing
  • Customer experience and customer care
  • Operations

These techniques can improve overall efficiency and agility by dramatically reducing detection times and the number of relevant events (e.g., alarms or trouble tickets).

The cloud, IoT and 5G era has begun. Service providers that adopt and master ML and AI for their big data will benefit the most. They will able to use this “new oil” to drive cost efficiencies, improve the customer experience and further automate network performance, scaling and security.

Sign up for the free insideBIGDATA newsletter.

Leave a Comment

*

Resource Links: