NFVgrid

Blend NFV Orchestration and Big Data-driven operations and analytics with NFVgrid.

NFVgrid brings together the revolutionary flexibility of NFV with the analytical power of Big Data methodology, all in one cloud-native package. With NFVgrid’s end-to-end network management and operations capability, provisioning and Day 2 functions become simpler, faster, and easier to control. NFVgrid is fuelled by traffic data itself to give you the clearest, most up-to-date picture of the network in action. That data also drives our proprietary, non-deterministic algorithms, enabling NFVgrid to apply machine learning to automatically adapt to network conditions on its own. And NFVgrid uses graph-based modelling for the topological representation of the hardware and network under-management in order to more accurately predict network load and configuration requirements.

What does this mean for your company?

Shrink development cycles and time to market. React faster to network load changes; predict them better – several times better. Seamlessly unite components from different manufacturers and even incorporate still-crucial legacy functions. Unlock the potential bound in siloed divisions. And say goodbye to costly bespoke network gear – NFVgrid runs on commercial-off-the-shelf server equipment.

NFVgrid organizes the revolution

Network function virtualization by itself has freed companies from a lot of specialized hardware and labor-intensive operations. But something needs to automate virtualization processes and interactions. That something is called an NFV Orchestrator. As an Orchestrator, NFVgrid takes any set of virtualizations and oversees their whole life-cycles and interactions. Whether the functions are completely cloud-based, or on bare-metal racks, or still on physical bespoke hardware, NFVGrid keeps everything talking to each other on a vendor agnostic basis.
In real life, the task set for an Orchestrator is even more complex. Companies need greater flexibility, better security, and finer-grained analysis to get the most out of their network operations.

NFVgrid, Big Data, and Orchestration

What differentiates NFVgrid from other Orchestrators is its cloud-native Big Data focus that fits the real world requirements of analytics-intensive companies. So, while NFVgrid handles Orchestration tasks such as the onboarding and management of network functions, it also enables extra-fine-grained data acquisition and analysis within and across those functions to determine what needs to be done in the first place.
We’ve taken Orchestration and made it smarter with analytical features that both determine and predict precisely what the network is doing – and what it will do next. NFVgrid features the network data analytics capacity companies need to drill down into the packet streams and then use Big-Data-based analyses to see what’s happening. Our proprietary algorithms enable you to predict your bandwidth needs in real-time, too. And NFVgrid’s network surveillance features let you can see with extra-fine granularity both exactly what the network is doing now and determine what it should do next, thus enhancing anomaly detection – and security.

Just how smart is NFVgrid?

How ‘intelligent’ an Orchestrator can be is a question on network operators’ minds. Here, there is no doubt: from smart discovery down to Tier2 through to machine learning, NFVgrid makes
Orchestration more than just an intellectual exercise. It finds out how the elements in the network interact and what the system’s normal operations look like. Then it troubleshoots whenever an anomaly arises or is about to, to handle what should be done, from onboarding further resources to raising an alert. In short, NFVgrid learns more, predicts more, and handles more.
Moreover, by providing layers of abstraction, NFVgrid makes it possible to look at a network with any level of granularity you require.

Let’s look under the hood

Architecture:

Virtualization is driven in part by the desire to unlock and break down silos, and NVF Orchestrators control functions across a variety of platforms. NVFgrid takes this principle one step further by being cloud-native and flexible.

The NVFgrid architecture is based upon four principles to ensure maximum compatibility:

  • Modular and extensible
    • Easy integration with any OSS, EMS, VNF Licensing manage
  • Open and Vendor-Agnostic with comprehensive networking support
    • Supports integration with any VNF, any OpenStack distribution, and any OSS using open APIs
    • Multi-Hypervisor support
    • Full Contrail integration
  • Telco-grade multi-site, multi-level deployment built to accommodate evolving standards
    • Supports multi-site deployments to ensure consistent management and common behavior of VNFs and network services
    • Full NFV orchestrator functionality, interfaces, and relational graphs
    • Can act as a VNF manager with complex VNF definitions or integrate if there is one
    • Bulk activation without using the GUI (API calls)
    • Inventory management and provision management tools
    • Host level DPDK support
    • VM level DPDK support
  • Consolidated and dynamic resource management at every level
    • VM segments
    • IP pool
    • VLANs/VxLANs
    • SDN flows
    • Ports, etc.
    • Predictive analytics (closed feedback loop)

Orchestration:

Picking up the cue from Cloud orchestration, NFVgrid sits in a separate layer between the VNFs and their managers on the one side, and Virtual Infrastructure Managers (VIMs), legacy network managers, and SDN controllers as well as physical infrastructure on the other. NFVgrid is in a position to onboard new functions, scale them up and down as needed, correlate both events and services, and terminate them neatly – and automatically, at that.

As a result, error rates drop, reaction times fall, and operator load (and costs) decrease. NFVgrid orchestration includes:

  • Network discovery and integrity
    • With NFVgrid, automating location-based discovery adds value across the network chain. Faster and more accurate identification not only enables resources to be used more efficiently, it also increases security and enables better security policy management. Combined with NFVgrid’s analytics capability, intrusion and unusual activity become much easier to counteract and, more importantly, prevent.
  • Automated provisioning
    • Automating provisioning is a central feature of NFV Orchestration for several reasons – dynamic servicing requires it and the whole concept of no-touch lifecycle management starts here. Getting automated provisioning right is critical, and NFVgrid’s analytics capabilities play an important part in making it all work. At the same time, instant placement optimization is based on sets of affinity and/or anti-affinity rules, geography, regulatory rules, etc., which are defined in NFVgrid’s easy-to-use GUI. There is also drag-and-drop provisioning and visual configuration of Virtual Network Appliances. UI-based network configuration includes:
      • Easy to use GUI for drag and drop provisioning
      • API-based bulk provisioning
      • Complete control in requesting infrastructure resources
      • Maintain inventory information
      • Predictive Analytics to eliminate over/under-provisioning
  • Topology Viewer
    • Visual representation of network topology with editing capabilities. Graphical representation in schematics and geographical view
  • Service Chaining
    • Assembling multiple virtual appliances into a single connected service chain with pre-configured setup for chained elements:
      • Software-configured service chaining to dynamically include best-of-breed functions
      • Highly dynamic or predefined service templates based service chains
      • Unidirectional, bi-directional and multi-directional service chains
      • Fine grained and coarse grained service chains
      • Pure or hybrid chains including Software Defined Networking (SDN), Network Function Virtualization (NFV ) and Service Function Chaining (SFC)
      • Automated and simple with flexibility (topology independent ), optimal (context awareness) and cost effective (on-demand resources) operations foremost
  • Lifecycle management
    • Provisioning, visual configuration, troubleshooting, firmware upgrade, retirement and other lifecycle management functions for VNFs
  • Product catalog
    • Simplified visual management of VNF for Provider’s service engineers to provision and manage to provide selected appliances to the end-user
      • NFV Service and VNF Catalog Management
      • VNF catalogue is a repository of all the on-boarded VNF Packages
      • It supports creation and management of the VNF Descriptors, software images, etc.; VNF packages
      • It is a repository of all usable VNFDs (VNF Descriptors)
      • The dashboard neatly presents all the virtual appliances and services that a customer can roll-out or terminate
  • Universal VNF manager
    • NFVgrid talks to the virtual appliances to push configurations with a defined on-boarding process for new vendors.

OPERATIONS:

From discovery to Day Two to life-cycle end, MFVgrid simplifies and automates operations. With a strong emphasis on daily operations and exception handling along with ease of use, NFVgrid helps improve service levels, speed up network reactions, and improve security while flattening human learning curves and reducing opex.

Operations includes:

  • Smart Discovery:
    • There’s more to network discovery than sweeping for devices. Along with finding the elements, NFVgrid determines the relationships between them. As a result, different levels of discovery are uncovered, including:
      • VNF discovery
      • Host level discovery
      • Service level discovery
      • Connection level discovery
      • Process level discovery
      • Network level discovery
      • Active/Passive actor based discovery
    • Relations between elements at any level are determined by packet flow information, object discovery and provisioning. Graph theory is implemented to understand the dynamics of those relationships in real time.
      Network discovery at this level helps administrators better handle events and alarms, and leads to better network behavior prediction.
  • Troubleshooting
    • Combining orchestration with Big Data techniques brings network troubleshooting to new levels. Troubleshooting becomes a multi-layered staple of lifecycle management and predictive provisioning. Basic troubleshooting and customer-facing diagnostics give clients a clearer view of perceived issues while enhanced status check and fault detection functions on the engineering side enable faster and smoother responses.
      • Service impact assessment
      • Fault localization
      • Identification of true root-cause
      • Taking corrective actions to resolve problems fast
      • Benchmarking tools (Generate User Load)
      • vPing
      • Define the troubleshooting in our own way (Temptest Scenario Tests )
      • Openstack Test Suite
      • OpenContrail Test Suite
      • OpenDaylight Test Suite
      • Reports (Generate Reports)
  • Machine Learning
    • Our proprietary machine learning algorithms merge data sources to drive more insights into how the infrastructure is performing. Machine learning is a critical part of the underpinning that advances NFVgrid’s anomaly detection features, which in turn ensures a more robust and secure network. This in turn is driven by our Big Data analytics, including:
      • Deterministic analytics
      • Network surveillance
      • Visual analytics: Instant Search and Visual Analytics based on a Collect, Store, and Analyze Everything model.
      • Predictive analytics: Guided troubleshooting. Predictive log analytics and machine learning self-learning, dynamic baseline and anomaly alerts
      • Operational intelligence: Business Value Dashboards (BVDs). capacity planning and optimization