Skip to content 
Organizations rely heavily on analytics to determine network health and predict future network requirements. The importance of network monitoring and automation through precise real-time analytics is widely recognized. On-premises analytics solutions provide useful insights into current network behavior and predict future issues based on historical data. However, these learnings are isolated to a single network. One of the advantages of considering a SaaS solution is its distinct advantage of learning not only from a current network but also from anonymized data collected from other similar ones. Ultimately, this deployment method provides a better understanding and prediction of overall network behavior through crowdsourced analytics.
Crowdsourced analytics is relatively new in the networking domain. However, its impact is widely recognized in other vertical markets. For example, Amazon product reviews leverage crowdsourced analytics to better recommend related products. Google Maps and Waze both improve mapping location accuracy through active user contributions. Consequently, it is no surprise that many organizations utilize crowdsourcing to improve accuracy and bolster security provisions to identify vulnerabilities and threats.
Crowdsourced analytics within networks can be leveraged to derive critical insights by collecting and processing information from various similar and dissimilar networks. Suggestions for SNMP or BGP configurations can be gathered by analyzing other networks belonging to the same domain or industry vertical. In contrast, security and VPN tunnel configuration policies specific to a location can be better understood by examining multi-domain networks in that geography
Benefits of Crowdsourced Analytics
Groups that are too much alike find it harder to keep learning, because each member is bringing less and less new information to the table. Homogeneous groups are great at doing what they do well, but they become progressively less able to investigate alternatives
The knowledge attained within an isolated network is limited. Siloed networks are slow to react to changing trends and fail to learn from other network operator mistakes. Sharing best practices and collaboration enables organizations to exchange new ideas and collectively learn and adapt network performance. From our perspective, applying crowdsourced analytics to existing network deployments provides four unique benefits.
Customized Suggestions
An essential function of any monitoring and analytics platform is to analyze the available data and provide suggestions to improve network health, performance, and availability. The quality of the suggestion depends heavily on the diversity of scrutinized data. Crowdsourced Analytics assimilates the diverse information gathered not just from the organization’s network but also from comparable networks in similar domains, locations, or scale to provide customized, accurate, and actionable recommendations. ATOM Cloud monitors essential data from existing networks such as throughput, scalability, device and interface utilization, errors and alarms, compares this data with anonymized data collected from many similar and diverse networks, and provides customized recommendations to improve performance and resiliency.
Network Optimization & Improvements
Network capacity planning and optimum utilization of network resources is an exhausting exercise. Even with a comprehensive plan and attention to the minutest details, optimizing the network to reduce OpEx is challenging. With crowdsourced analytics, ATOM Cloud can correlate network data such as neighboring flaps, packet drop, data latency, QoS, and other variables to identify and measure what operators should undertake to minimize resource utilization fluctuations. By receiving suggestions on a similar network’s performance in real-time, operators can predict traffic in their network and better allocate related resources..
Predictive Assurance
Maintaining network baseline behavior is challenging. Network health information from the analytics engine is conveyed to the automation engine to take appropriate provisioning or remediation actions. Even a small deviation from baseline can potentially lead to significant changes in routing, tunneling, and related security configurations. To ensure strict SLAs, the network must identify possible issues ahead of time. Using AIOPs, crowdsourced analytics helps the assurance engine identify similar problems in other networks and use that learning to predict and enforce changes to improve network stability and performance. Through predictive assurance, ATOM Cloud can investigate network traffic and operations to predict the existence of active DDOS threat, device & system malfunctions, or abnormal spikes in demand.
Acceleration of Time to Market
Another fundamental benefit of crowdsourced analytics is improved service delivery. Developing, testing, and validating every policy is time-consuming and tedious. Crowdsourcing allows organizations to unleash the power of the community. Crowdsourcing also enables the discovery of solutions to common problems through intensive collaboration. Organizations can leverage various analytics or assurance scripts developed by their peers, such as service provisioning, threat analysis and prevention, proactive actions on suspicious behavior, and rerouting alerts based on severity and correlation. As a result, network operators can easily modify scripts to their needs and speed networking deployment. ATOM Cloud leverages crowdsourcing to accelerate service delivery, reduce OpEx, and increase return on investment.
Advantages of ATOM Crowdsourced Analytics
ATOM Cloud leverages various crowdsourcing technologies to provide enhanced network automation and assurance solutions that deliver the following unique benefits:
Security and Privacy
ATOM Cloud collects anonymized data, normalizes and processes it, and makes it available to all instances. The analytics and assurance engine uses learned network data, compares it with existing data and uses it in complex prediction algorithms to assess and forecast the network’s state. ATOM Cloud is equipped with RBAC, SSO, multi-tenancy, integration to TACAS, and various other security features to prevent unauthorized access and distribution of sensitive information.
Distributed Low-Latency Architecture
For analytics to be effective, the data must be analyzed as close to real-time as possible. ATOM Cloud provides a distributed architecture to collect, visualize, and process data with minimal latency. ATOM Cloud collectors can also be deployed at the edge of the cloud or on-premise for the ultimate degree of flexibility.
Multi-Vendor and Multi-Domain Capable
IT networks consist of connected devices from multiple vendors. Therefore, to provide a genuine crowdsourced capability, any analytics solution must be multi-vendor. A significant advantage of crowdsourced analytics is its ability to provide relevant suggestions from different networks. Thus, the solution must also be multi-domain capable. ATOM Cloud supports more than 45 vendors across campus, branch, service providers, and enterprise domains. The platform’s vast multi-domain and multi-vendor capabilities also allow the ATOM Cloud analytics engine to collect approved and anonymized network and device information from thousands of devices across hundreds of customers, to develop enhanced network insights.
Intelligent Learning
ATOM Cloud utilizes artificial intelligence (AI) and machine learning (ML) to correlate large volumes of data received from various sources. AI/ ML capabilities are employed to provide customized suggestions, alert and event suppressions, root cause identification, related automation and security policies, and an improved user experience. Intelligent learning forms the heart of all ATOM Cloud operations.
ATOM Cloud is spearheading the network assurance and analytics revolution. It enables organizations to assess and improve network policies through customized recommendations and automated remediations powered by collaborative learning.
To learn more about ATOM Cloud, visit here.
