2. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
Ari Banerjee
VP Strategy
Netcracker
James Crawshaw
Senior Analyst
Heavy Reading
Martin Thygesen
Crosswork Product Manager
Cisco
Juan Manuel Caro
Operations & CEX
Global Director
Telefonica
Panel
Roy Silon
VP R&D
Atrinet
3. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
• Operators want analytics tools to provide them
with tangible insights: findings that are actionable,
concrete, palpable.
• At the same time they want these systems to be
highly automated, employ artificial intelligence
and be zero-touch.
• So, palpable and zero-touch at the same time;
quite a challenge.
Introduction
4. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
1. Create data repositories (centralized and distributed)
2. Select sources of data (contact center calls, field
technician reports, bills, energy usage, OSS, network
telemetry, etc.)
3. Anonymization (GDPR, etc.)
4. Normalization using a standard data model
5. Check data quality (inventory, data capture/entry
processes)
6. Analyze use cases
• Operations related (e.g. infrastructure management,
customer experience, customer service delivery, internal
plant management, etc.)
• Technology related (video platform, radio planning, etc.)
The Data Analytics Process
5. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
“The best strategy for data lakes is to only
collect data that is useful now. Data loses its
value over time and if you can’t find what
you’re looking for in the mess that is the
data swamp, it’s pointless to keep adding to
it. Projects should only go after sources that
can provide useful solutions to clearly
defined business problems.” - Nick Ismail
https://www.information-age.com/dont-drown-data-lake-123466667/
Don’t Drown in a Data Lake
6. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
• Big data technologies enable CSPs to store vast
quantities of data about their networks and services …
• … but is filling huge data lakes with streaming
telemetry about network paths, traffic flows and
performance going to provide a valuable resource for
analytics or simply rack up a big bill for storage?
• Might CSPs be better off applying a coarser filter to the
data they collect, focusing on the metrics which are
likely to have a material impact on performance?
• Do the IT development team tasked with building the
data analytics platforms have enough networking
expertise to know which data is worth keeping?
How Big Does Your Data Need to Be?
7. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
Use case Description
Network
proactive
maintenance
In some scenarios, Telefónica can now predict outages in network equipment, so that all
teams in operations can focus on avoiding the impact on services and customers.
Instead of waiting for faults to happen, our network teams can know in advance what is
likely to happen and analyze the possible impact. With this use case Telefónica is
improving network availability by 0.2pp.
Predicting
outages in
submarine
cables
Telxius is now able to predict when some important weather conditions (hurricanes,
tsunamis) are going to affect connections through its submarine cables. This information
allows them to proactively change routes and avoid giving bad service to their
customers. Also they are more efficient because they can buy capacity from external
providers with cheaper costs.
On-site works
prioritization
Traditionally on-site works are done in a first-in/first-out scheme. This could lead to
situations when problems in the network affecting many customers are not properly
prioritized. Now, our algorithms can work with different criteria and recommend new
prioritization schemes that focus mainly on customer impact. As a result, network
availability and customer satisfaction have improved.
Network
CAPEX
allocation
Our algorithms allow us to prioritize our investments assuring our customers the best
service possible. From new schemes to locate spare parts to minimize outages or to
deploy new network equipment, to new ways of assigning works to contractors to ensure
the best quality and prices.
Telefonica Case Study – Use Cases
8. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
Incorporating AI into Analytics
• Most analytics use traditional statistics
• AI could enable more intelligent, predictive and
adaptive analytics
• Next best action for staff in the SOC
• Orchestration
• Video anomaly detection
• Real time index of customer satisfaction
9. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
• Could CSPs open up their anonymized data sets to
third parties so that they can propose their own
uses cases and analysis?
• CSPs could share a portion of any value found with
the third party.
• How would you audit this?
Crowdsourcing CSP Data Analytics
10. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
How would you describe your end-to-end network
understanding, based on network discovery?
0%
17%
21%
62%
We are not using discovery at all, as we don’t find
it crucial to our processes
Discovery based on our existing OSS is sufficient to
understand my existing network and planning
future needs
Current inventory is inaccurate – we need a more
accurate network discovery (but not necessary
real-time)
We need a network discovery that is 100%
accurate, based on real network data, and in
real-time
Source: Light Reading webinar survey 6 June 2018, n=62
Data Accuracy
11. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
Which of the following best fits your view?
• Data analytics should be a core competence of
the CSP
• We already have plenty of data analytics – what I
need are actionable insights
• Data analytics is over-hyped – I’m running the
network just fine, thank you
To vote please go to pollev.com/jamescrawsha760
Audience Question Poll
12. P R E S E N T E D B Y :
@LIGHT_READING #CNGEUROPE
Ari Banerjee
VP Strategy
Netcracker
James Crawshaw
Senior Analyst
Heavy Reading
Martin Thygesen
Crosswork Product Manager
Cisco
Juan Manuel Caro
Operations & CEX
Global Director
Telefonica
Panel
Roy Silon
VP R&D
Atrinet