Uhana by VMware

Real-time RAN Analytics for

Mobile Operators

AI-based Analytics for Mobile Networks At-a-Glance

Uhana by VMware is an AI-based analytics platform that provides • Subscriber-level RAN
mobile network operators (MNOs) the ability to generate actionable analytics in real time for
insights from their 4G & 5G radio access network to improve new insights
customer experience and reduce operational costs. Uhana correlates
• Improve customer
and enriches high volumes of streaming data from mobile network
experience by
elements in real-time to provide unprecedented visibility at a
understanding call quality
subscriber-level. Using artificial intelligence, Uhana automatically
and data performance on a
determines the root cause of network issues and recommends
per subscriber basis
remediations.
• Prioritize alerts based on
subscriber and network
impact

• Eliminate ‘average
blindness’ with granular
RAN visibility

• Use Machine Learning to

detect, classify and localize
issues in the mobile network
in an automated fashion

• Automate Root Cause

Today’s Reality Analysis and recommend
Mobile network operators are limited by the state of traditional radio remediations via AI
access network (RAN) analytics and telemetry. Coarse-grained • Automate data preparation
telemetry data presented at a cell site level, in 15 minute increments, for data science analysis and
severely restricts performance, customer experience measurements realize time savings of 30%
and radio interference detection to historical analysis. Because actual
network conditions change within seconds, any user experience or
network efficiency guidance derived from coarse-grained data will
suffer from “average blindness” which severely limits visibility. As 5G
latency-sensitive applications are being rolled out and network slicing
becomes mainstream, this historical, ‘average’ analysis becomes
crippling and will hinder any meaningful, real-time corrective actions
in the RAN.

Datasheet 1
Uhana by VMware Innovative Approach Benefits

Uhana by VMware has taken an innovative, automated approach. Fine-Grain Visibility

Leveraging machine learning and AI, Uhana detects, classifies and • Subscriber-level
localizes issues in the mobile network in an automated fashion. The • Sub-second granularity
Uhana AI platform ingests and processes concurrent call trace events
from tens of thousands of cells, correlates this with user session data Automated Data
and calculates Key Performance Indicators (KPIs) in real-time. Mobile Enrichment Pipeline
network operators can now see detailed subscriber session records • Joins call trace events
within seconds. This granularity gives operators unprecedented • Creates subscriber
visibility and allows them to see problems that were previously hidden. session record

Automated Incident Detection

• Detects anomalies
• Leverages neural networks

Cloud-native

Uhana by VMware is a state-of-the-art cloud-native application

leveraging a micro-services architecture. Built on Docker containers,
micro-services enable resiliency and scale-out and allows the platform
to be deployed on bare metal or virtual machine infrastructure.

Automated Incident Detection and Impact Prioritization

Uhana’s AI-enabled Alerts Pipeline automatically establishes dynamic
baselines for subscriber level KPIs that adjust according to the ebbs
and flows of the network. Using dynamic thresholding, it then detects
anomalies, clusters them using neural networks and generates alerts
as needed. Alerts are prioritized based on impact, for example,
how many subscribers are affected and the significance of service
degradation, so that responses can be prioritized accordingly.

Datasheet 2
AI-driven Root Cause Analysis Benefits, Continued
Each alert is further analyzed by identifying a root cause (leveraging
neural network classifiers) and providing relevant KPIs and AI-based Root Cause Analysis
recommendations specific to the identified root cause. Operators no • Automatically determines
longer have to spend precious engineering resources on analysis to root cause
identify the true cause of the problem. In addition to enabling quicker • Recommends remediations
remediation actions and better control, correctly diagnosing the root
cause of issues enables predictive network intelligence for application Cloud-native
and RAN control guidance. • Containers-based
• Micro-services architecture
.
Carrier Scale
• High velocity (millions of
events per sec)
• High volume (tens of
thousands of cells)

Use Cases
All use cases leverage AI & ML for detection, impact and ‘what-if’
analysis and guidance.

Uplink Interference Detection

Automate detection and classification of uplink interference

(external, PIM, infrastructure) with prioritization based on type
and number of subscribers impacted.

Interference Localization and Triangulation

Localize external interference with a prioritized list of search

area and heatmap visualization. Reduce area of interference
triangulation from miles/kms to blocks, saving hours of
reconnaissance time.

Datasheet 3
 Downlink Throughput Impact Benefits, Continued
Customizable
• Design custom KPIs
Use machine learning root cause and impact analysis • Deploy in real time
algorithms to determine if cause of poor downlink throughput is
related to load imbalance of antenna frequencies for optimum use Fast Time-to-Value
of spectrum. • Offered as a managed SaaS
• On-prem or in public cloud
CAPEX Prioritization
Data Science Friendly
• Automates data
Increase effective spectral capacity by improving RAN enhancement
optimization engineering. Prioritize network investments that • Reduces time spent by 30%
maximize subscriber impact and reduce CAPEX spending.

OPEX Reduction

Automate RAN optimization analysis and improve efficiency of

RAN optimization engineering & operations through AI based
RCA, granular visibility, and impact prediction.

Automated Data Enhancement for Data Scientists

Reduce data science analysis cycle time by up to 30% through

the automation of data preparation, freeing up valuable time for
higher value functions.

Uhana Building Blocks

The Uhana AI platform is an open multi-stage stream processing
AI pipeline that operates on an operator’s cloud or public cloud
infrastructure. It is managed by VMware and offered as a Managed
Software as a Service (SaaS). The platform leverages RAN and
evolved packet core (EPC) telemetry information streamed directly
into the data ingestion and distribution engine from the RAN
infrastructure and EPC. This data is not inline to the user plane and
therefore does not affect the performance of the network. This data
is then processed by the real-time stream processing engine, which
calculates relevant KPIs and feeds the Uhana AI-engine for learning
and neural network creation. This ultimately provides guidance to
operators via the Uhana API which can be applied manually or directly
through closed-loop automation.

Datasheet 4
Features and Components
Data Ingestion and Distribution Engine
Ingests real-time telemetry information from a variety of sources,
such as cell trace records (CTR) from eNodeBs, and cell trace
UE-ID mapping (CTUM) from MMEs, and replicates it to multiple
downstream systems via an open API.
• Scale-out ingestion of real-time encoded RAN and
EPC Telemetry
• Data ingestion (1M events/sec) and replication (3M events/sec)
• Support for multiple simultaneous RAN software releases for
in-service RAN/EPC release upgrades
• Real-time feed replication for up to 3M events/sec
• Ability to replicate full/partial feed based on
pre-filtering criteria
• Dynamic Kafka topic management with authentication
• Optional anonymization of selective fields (IMSI, IMEI, etc.)

Datasheet 5
Real-time Stream Processing Engine
Uses real-time feeds and applies stream processing, for example:
data filtering, joining and KPI calculation. Over 400 KPIs are
supported out-of-the-box and custom KPIs are defined via the KPI
composer and calculated in real-time. The resulting streaming data is
accessible via the Uhana web dashboard, the Uhana dashboard API
and Kafka-based open API. Data can be sliced in multiple dimensions
(e.g. per-QCI, per device type, per device release, etc.), aggregated
over different time intervals, and calculated using different
statistical functions.
• Real-time stream processing for RAN KPIs (with resolution as
high as 10 seconds)
• Sub-second processing latency at scale
• Access to real-time and historic operations data with
configurable data retention policy
• Rich filtering capabilities (e.g. per-UE, UE group, QCI, device
type, device release, bands, sectors, geographic regions,
antenna type, etc.)
• Temporal KPI aggregation
• Dynamic threshold based alerting

Network Operations Optimization and Automation Suite

Consists of applications focused on optimizing mobile operator
infrastructure. It applies Uhana’s AI and ML technology to detect
network anomalies, predict network KPIs, offer control guidance
or provide closed loop control automation. This suite is made up of
independently licensed applications that can range from anomaly Learn More
detection, interference detection, end-to-end service assurance
For additional information
and more.
about Uhana by VMware:

1-877-VMWARE
(outside North America,
dial +1-650-427-5000)

telco.vmware.com

VMware, Inc. 3401 Hillview Avenue Palo Alto CA 94304 USA | Tel 877-486-9273 | Fax 650-427-5001 | www.vmware.com
Copyright © 2019 VMware, Inc. All rights reserved. This product is protected by U.S. and international copyright and intellectual property laws. VMware products are covered
by one or more patents listed at http://www.vmware.com/go/patents. VMware is a registered trademark or trademark of VMware, Inc. and its subsidiaries in the United
States and other jurisdictions. All other marks and names mentioned herein may be trademarks of their respective companies.
6