中文
5G is ushering in a new era for telecom, promising literal life-changing capabilities built on ultra-fast and ultra-reliable services that are adaptable and dynamic. By embracing new approaches such as cloud-based networking, multi-vendor environments, private networks, edge computing, and more, 5G promises a quantum leap forward in delivery speeds and reliability. But this revolution in service delivery will require a substantial evolution in assurance approaches if providers are to guarantee the promises of 5G are delivered.
Over the past two decades, providers have spent hundreds of millions of dollars instrumenting the network with traditional passive probing assurance systems — yet still struggle to isolate the root causes of end-to-end issues, avoid major network outages, and automate problem resolution. The introduction of 5G has exponentially compounded these difficulties as the underlying network infrastructure is now fully virtualized, highly disaggregated, and oftentimes only partially owned by the provider. Next-gen 5G networks are much more dynamic than traditional networks, changing so often that fixed instrumentation deployments reliant on actual subscriber events happening is quickly becoming a relic of a bygone era.
Two main types of assurance approaches
Assurance methods are typically either considered passive or active:
Passive assurance: Physical probes are installed at key points in the network to collect and analyze real user traffic flowing through those nodes. Aggregation of this data is used to generate key performance indicators (KPIs) that, in turn, help to detect problems and are used for the basis of setting and triggering alarms and alerts.
Active assurance: Virtual (or active) test agents are integrated into and throughout the network infrastructure. These virtual points of presence are used to generate small amounts of emulated user traffic. This deterministic data set is then aggregated in support of KPIs and is also used to segment and isolate issues across complex networks.
Most articles you read on “passive versus active” assurance promote a dual approach predicated on different use models. Historically, active testing has been relegated to pre-deployment use models or transport layer health checks; whereas passive probing has been the predominant paradigm as it relates to service (application) assurance. 5G has created an inflection point and accelerated the market’s interest in utilizing active testing in an always-on service assurance model given its flexibility and extensibility within dynamic environments. The answer to the question of why any provider wouldn’t forklift their passive probing infrastructure in favor of an active test deployment are both commercial and technical.
Historically, active testing has been relegated to pre-deployment use models or transport layer health checks ... 5G has created an inflection point and accelerated the market’s interest in utilizing active testing in an always-on service assurance model given its flexibility and extensibility within dynamic environments.
Challenges for active testing
Commercial challenges
Passive probing: Represents years of embedded deployment and support that are defended by an incumbent who wants assurance methods to remain status quo. While these individuals may not be the budget owners, they exert great influence over buying decisions.
Active testing: Being the “new kid on the block” means that any continuous active test approach must prove itself as not only capable but better than the long-standing alternative. This is represented in extended POCs and trials fronted with a “prove it to me” mentality.
Technical challenges
Passive probing: Represents a largely independent system that is not reliant on deep-seeded knowledge of the underlying network infrastructure. A basic understanding of the speeds and feeds in support of the probing interface will often suffice.
Active testing: Can require a more intimate knowledge of the underlying infrastructure, given the test agents and emulated traffic are a functioning part of the network vs. an external device observing the network.
Both the commercial and technical challenges of converting from passive to active are made exponentially easier with the 5G tech turn. Rarely does a quarter go by that you don’t read about a major provider outage who is working through a 5G upgrade. All of which have substantial passive probing deployments that are proving less useful every day.
Key differences with passive vs. active assurance
Monitoring and implementation
Passive assurance: Monitors specific high-traffic aggregation points in the network; on one end or the other in the deployment.
Active assurance: Measures performance anywhere on the network using emulated traffic and can be located at either end or any point in between.
Performance measurement
Passive assurance: Measures performance only when the network is active with live customer traffic
Active assurance: Can measure performance with simulated traffic before users are live on the network (for pre-deployment validation).
Adaptation
Passive assurance: Static positions of hardware appliances do not adapt to topology changes in dynamic networks.
Active assurance: Automatically mimics the service as networks change to optimize itself without interrupting KPI generation.
Issue identification
Passive assurance: Detects major issues and determines how many users are impacted.
Active assurance: Identifies minor problems before they become major, customer-impacting issues, avoiding SLA violations.
In 4G LTE, the core network would typically consist of no more than four datacenters with 8-10 critical ingress/egress points to be probed passively. In a native 5G deployment, the provider would have to deploy and support hundreds of passive probes.
Key benefits of active assurance
To recap, in a 5G world, the benefits of implementing active assurance throughout the network are many:
Streamline activation testing. Activate new network functions and services with emulated user traffic to predict performance before rolling out to live operations.
Automate troubleshooting. Machine-learning detects irregularities and automatically initiates troubleshooting test procedures.
Deliver reliable performance. Avoid costly SLA violations by proactively identifying and remediating potential issues before services are disrupted.
Proactively detect issues. Verify performance before usage starts, when traffic levels are low, to uncover minor issues before they have a major impact.
Gain end-to-end visibility. Use complete network and service visibility (end-to-end and within each domain) to quickly pinpoint issue domains for faster remediation.
Leverage synthetic traffic. Stop waiting for an actual customer-impacting event to identify performance issues.
To learn more about how active assurance can promote confidence in network performance, read the Spirent eBook, 5G Active Assurance: Embracing a New Perspective.
