Organizations now can take real-time, proactive control to remediate issues and protect against network threats with software-controlled sensors and data plane engineering capabilities.
How to Manage Real-Time Intelligence at the Edge
In this article, you’ll learn:
- The drivers behind next-generation networks
- The importance of analytics at the edge
- The role of MantisNet’s solutions in helping manage real-time intelligence at the edge
Content Summary
Next-Generation Networks – 5G & IoT
The Importance of Analytics at the Edge
How to Manage Real-Time Intelligence at the Edge
Location, Location, Location
The Rising Importance of a Programmable Data Plane for 5G and IoT Success
Deploying continuous real-time intelligence at the edge to: utilize data plane engineering capabilities to monitor, manage, optimize, and control traffic while increasing the performance, reliability, and security of next-generation networks.
Next-Generation Networks – 5G & IoT
Next-generation networks and IoT applications are expanding rapidly due to the introduction of 5G connectivity and IoT devices. These new technologies will transform all facets of life in nearly every industry. Applications ranging from smart grids, virtual power plants, smart homes, intelligent transportation, next-generation industrial control systems, and smart cities will expose organizations to a wide variety of new vulnerabilities and attack vectors resulting in a myriad of new operational and security challenges.
With an exponential increase in the number of access points, the edge of these networks must be optimized to be able to filter, slice, identify and process both existing and new traffic types at all layers as well as to continuously deliver telemetry in real-time, to better react to these new attack vectors and performance obstacles. Additionally, businesses will demand real-time interaction with their services and customers – making continuous real-time monitoring and management of data essential as computing and processing shift further to the edge.
According to a 2017 Boston Consulting Group report, the market for IoT products and services is expected to reach $267 billion by 2020.
Gartner, Inc. estimates that 20.4 billion IoTconnected components will be connected worldwide by 2020, and more than half of major new business systems and processes will include an IoT component.
Looking ahead, next-generation networks will increasingly rely on hybrid Software-Defined Networks (SDN) and advanced analytics. Specifically, SDN tools working in conjunction (AI and ML) analytic workflows that interact with new and traditional network infrastructure, enabling more powerful network management and control capabilities. The result is that the edge of the network becomes a more critical surface for data interaction.
The Importance of Analytics at the Edge
With billions of devices potentially connecting to the internet, the only way to effectively manage and control complex network topologies and the sheer volume of traffic is to move the intelligence closer to where the data is generated. Consequently, the most critical systems will be the real-time intelligent sensor technologies embedded where traffic from those devices enters the network – at the IoT gateway. An IoT gateway aggregates, filters, and processes network traffic and produces telemetry and sensor data using embedded instrumentation and transcoding the various device and network protocols while working with analytics to pre-process network traffic before forwarding it. According to Enterprise Management Associates’ (EMA) research, the most significant function of IoT gateways is the support of edge computing and analytics.
MantisNet sensors and software technologies can add significant value in the deployment of next-generation 5G networks and the overall IoT technology stack. MantisNet software sensor (transcoder) technologies can be embedded anywhere in the network; at the core, mid-span at the edge, or embedded in the IoT gateway, to serve as the vital sensing component to “feed” continuous, real-time event stream processing and advanced analytics (machine learning and artificial intelligence) capabilities required to keep up with the demands of IoT and next-generation networks. The continuous monitoring and real-time remediation capabilities of MantisNet software sensor technologies also help to improve the service lifecycle as well as lowering costs by optimizing traffic flows across new and legacy infrastructure.
Our premise is—with the rollout of next-generation networks and IoT devices—existing forms of centralized monitoring and analytics will not scale due to the sheer amount of data, the distances involved, and the short decision loops that will be required to maintain and protect those systems. The best way to effectively monitor and manage these systems—as well as to deliver real-time detection, decisioning and response capabilities—is to decentralize monitoring and event processing by placing sensor and traffic engineering functions closer to the edge, which can be achieved using intelligent MantisNet sensors combined with embedded event stream processing analytics. This aligns with the prevailing industry position (per Cisco, Intel, et al.) that distributed, “analytics at the edge” are critical to the successful deployment of these systems.
How to Manage Real-Time Intelligence at the Edge
MantisNet sensor technology—deployed at the edge, in the cloud or the data path—is uniquely capable of facilitating the delivery of continuous monitoring and responsive analytics by transcoding, extracting, and generating new and unique forms of high resolution, high-performance telemetry in the form of streaming metadata.
MantisNet’s programmable sensor technology consists of P4 software running on SDN infrastructure, providing software-controlled traffic shaping, processing, instrumentation, monitoring, and visibility across any network without the need to replace existing infrastructure or deploy separate management and monitoring equipment. MantisNet systems provide advanced instrumentation, high-resolution visibility, traffic management, and improved operational efficiencies for existing networks and the next generation of 5G-driven IoT infrastructure.
Location, Location, Location
By placing MantisNet’s dynamically programmable sensor and data plane management technology close to, or where the data is generated, the systems can dramatically improve the accuracy, efficacy, efficiency, and speed of artificial intelligence (AI) and machine learning (ML) IoT analytics. With advanced processing of network traffic where/when it is created, MantisNet allows for improved:
- Connectivity and security: monitor and ensure the security and integrity of the network and network-attached devices
- Traffic visibility: decode and transcode multiple protocols and data formats in real-time into high-resolution metadata that can be used by localized or centralized monitoring or analytics systems
- Management: streamline the ability to provision, update and control IoT devices, network traffic as well as apply policy-based permissions
The Rising Importance of a Programmable Data Plane for 5G and IoT Success
Moving sensor technology to the edge- where the data is generated- is only one component of a strategy for successfully deploying 5G and IoT networks. These systems should also have data plane programming capabilities and application programming interfaces (APIs) to enable organizations to dynamically interact with a data plane to better control the flow of information moving over their networks.
Data plane interaction is key when it comes to deploying next-generation networks, particularly when considering high-speed, complex, application-driven networks. The data plane is where data flows – it consists of the payload (data) and all the control information – the actual substance of the information traveling throughout your networks. Over the years most network equipments relied on proprietary control plane interactions. More recently there has been the emergence of intelligent software-defined networking (SDN) solutions that can manage a wide variety of switching and routing hardware. With the emergence of these SDN capabilities vendor lock-in has begun to erode, and more meaningful work is being done by network engineers who understand the power of programmable networks to architect solutions that best fit their environments.
Data plane engineering has only recently emerged in the networking space. SDN technology applied at the data plane enables these capabilities by providing the ability to have dynamic, programmatic interactions with the data. Solutions using programmable data planes and sensor technologies from MantisNet have begun to change this conversation enabling the complete control and flexibility required for next-generation networks.
Consider the basic components of network monitoring – you have the network, from which you filter or extract data for analysis, and you have the analytic layer- the brains behind the operation, driving all decisions. With MantisNet programmable sensor technologies, organizations can get real-time visibility into events as they occur and the ability to interact with the data plane to ensure that the most meaningful information is available to the analytic layer. Combined with network traffic engineering, organizations can process network traffic at wire-speed with in-memory sensor technology – programming them to extract meaningful metrics to feed analytic workflows as well as manipulate the data plane to perform dynamic traffic shaping. Deep insights into the entire protocol stack (L2 – L7) can also be normalized through pre-processing and served up into streaming analytic workflows in the form of streaming metadata. Furthermore, because of their programmatic nature, the sensors can accommodate all existing as well as any new, or emerging, protocols as they develop.
Organizations now can take real-time, proactive control to remediate issues and protect against threats with software-controlled sensors and data plane engineering capabilities. Network information can be extracted and processed for analytic consumption, the analytics themselves can interact with the sensors to identify the bottlenecks, failures, malformed packets, or cyber threats, etc. that need to be remediated. Using data plane engineering provides a wide range of capabilities including rule-based flow masking, deep packet filtering, dynamic parsing / de-parsing, load balancing, traffic shunting, and rerouting traffic in real-time. By using software communications capabilities (APIs), the integrated sensor and data plane engineering systems working with the analytic layer, to dynamically interact with the data plane to apply changes to the network in real-time.
MantisNet solutions are at the forefront of expanding the capabilities of SDN technologies. MantisNet solutions excel at all these capabilities and our experiences cover the gamut from continuously monitoring and decoding any/all known or new network protocols to providing new and unique forms of proactive authentication and fraud detection technologies. MantisNet solutions are coupled with our deep experience in real-time network monitoring and can be relied on to deliver the management and control that next-generation networks and IoT applications will demand.
Looking ahead, MantisNet will continue to extend the capabilities of real-time monitoring and data plane engineering solutions to further enhance the next generation networks and IoT ecosystem with innovative, scalable solutions. MantisNet is committed to solving the most difficult challenges and delivering compelling new capabilities for the current and next generation of networks.
Source: MantisNet
