5. INTERNAL
5
Renewable capacities (GW)
Users (mln)
2021 2024 2030
SAIDI (min)
Renewable production
(TWh | share%)
Zero emissions production
(incl. nuclear)
RAB (Mld€)
ENEL Group : The transformation of the Group
Demand Response (GW)
Electric Bus (k)
Battery Storage1
(MW)
Charging points (mln)
Electricity sold 2 (TWh)
1. Behind the meter
2. Includes power customers - free and regulated market
53,4
118 | 51%
62%
75,2
243
43
~186 | 67%
77%
81
216
~77
49
~340 | 80%
>85%
86
~ 100
65
~154
309
7,7
>3
80
0,32
~13
13
~300
1,1
~353 ~550
>20
>20
>1.000
>5
6. INTERNAL
6
ENEL Group : The energy transition is driven by …
Urbanization
New customers’ needs
Digitalization
Electrification
Decarbonization
Energy Transition is not a new trend, it is just happening faster
7. INTERNAL
7
ENEL Group : Towards a platform company
Our
vision
Customer
Management
Buy
Sell
Unique database
& Network Digital
Twin
Global
Activities
Asset Operator
Asset Owner
Physical
assets
Digital
layer
Customer identity
Standardized
Back Office
Front
End
Back
End
Products
Distribution Enel X
Retail
Customers
Suppliers &
partners
Cities
Prosumers
Assets (e.g.
charging
stations)
Source: Enel Capital Market Day Strategic Plan 20-22
8. INTERNAL
8
Technological
architecture
Work organization
Reskilling
Business solutions
Processes
• Access to data across whole organization, with no duplication of entry points
• Open platform enabling quick response to internal and external stakeholders
• Implementation of data driven decisions across all the value chain
• New approach based on 3 principles:
– Simplification
– Streamlining
– Rationalization
• New processes mapped in terms of capability needs
• Needs for new profiles and skills
• Data driven business solutions to enhance automation, assets monitoring and
customer experience
• E2E clean-sheet redesign aiming at effectivness and efficiency optimizing: digitalization,
Ins/Outsourcing mix, capabilities
• Modular and integrated framework to handle complex and evolving environment
ENEL DH I&N : GBS Program
10. INTERNAL
10
Project Needs
Entity Access (typical SQL login).
Good performance in accesses based on
plant properties, in order to retrieve the
attributes of a node without considering
the relationships.
Need for a db that could easily represent the model of
the electrical network.The graphs represent this
infrastructure well.
Navigation of the relations of a graph (typical
graphDB access). High performance in accessing
network objects related to each other, in order to
extract entire power lines and all the elements that
compose them
11. INTERNAL
11
3 candidates
• Neo4j
• Other graph db
• SQL DB
3 databases of
different sizes
• "Small" DB containing
the network of a country
• «Medium» DB containing
the network of 3
countries
• "Large" DB containing
the 3 country network
multiplied by 5
21 sample graphs
Starting from the three
types of network (T1 / P2
/ S2), 21 sample graphs
representing parts of the
network were extracted,
chosen by number of
elements and depth
Decision making process
12. INTERNAL
12
The advantages of a Graph DB - Why Neo4j
Feature Neo4J Other graph DB SQL DB
Scalability Horizontal Vertical (no sharding) Vertical (no sharding)
Availability
- Master-Slave data replication
- Supports full / incremental
backups from the running cluster
- Monitoring and restoring
instances
- Supports up to 15 read replicas
ACID Compliance Yes [1]
Yes, except for some operations in
order to increase performance
Yes
Supported graph
models
Property graph [Cypher e Gremlin]
- Property graph [Gremlin]
- RDF [SPARQL]
None [2]
Data visualization
Neo4j Bloom, integrated and highly
customizable
Absent, but third-party solutions
that can be integrated
No graph data display support
Security
- User role management
(Enterprise ed.)
- Support external authentication
systems via LDAP (eg Kerberos)
- Access management to portions
of graphs
- Isolation in VPC
- Permissions managed through
AWS IAM
- Cryptable instances with AWS
KMS
- Automatic update management
Fine grained access rights
according to SQL-standard
Graph data
analysis
Neo4j Graph Analytics, library
composed of procedures that can be
called up by cypher (centrality,
clustering, pathfinding)
Not supported Not supported
Ability to edit Yes by modifying the vertex labels
Yes through workaround on the
creation of new vertexes with
updated labels and elimination of
the old ones
Yes
License Open Source/Commercial Commercial Open Source
neo4j Other graph
DB
SQL DB
* Features analyzed at the date of the POC(Q4 2020)
13. INTERNAL
13
≈ 100 µs
• 500k queries / day
• 10 countries
10 DBs (1 per country)
• 1TB data
• 600M nodes
• 800M relations
For each environment (Dev, UAT, Prod)
• 3 server Causal cluster
• 128GB RAM, 16 cpu
Implementation process and results obtained
15. INTERNAL
15
- Guarantee the availability of
the 24x7 service with
dedicated support
- Support high concurrency
parallel logins
- Support parallel massive
extractions
- Ensure high data access
performance
- Maximize high reliability
Next steps - New challenges for the future