NECOS Industrial Workshop Technical highlights by Prof. Alex Galis (University College London – UCL)

NECOS Project Technical Highlights
http://www.h2020-necos.eu
Industrial Workshop
CPQD Tech Park Campinas
18th October 2019
1
Prof. Alex Galis
University College London
a.galis@ucl.ac.uk
www.ee.ucl.ac.uk

Agenda
• https://youtu.be/8l9mzVf48oM
• Context and Motivation
• NECOS Approach: Lightweight Slice Defined Cloud (LSDC)
• Results & Achievements
• Key Value Proposition & Concluding Remarks
2

Networking Context & Trends
3
L1-L5SpecializedHardware
L5-L7ClosedH/W
withembeddedS/W
L2 - L7
Software:
Components
and platforms
on
General
purpose
Hardware
L1 Hardware
2. Softwarization
3. Dynamic Interaction between groups of communication, compute, storage and network
services/applications elements/devices in all network segments (edge, core, wire/wireless access, space)
4. Cross Layers new requirements /characteristics: different and very stringent non-functional requirements
including the strict low latency and high data exchange requirements and guaranties for KPIs and/or SLA
characteristics per parts of the infrastructure (SLICES).
1. 5G+ is an integrated, highly automated and intelligent infrastructure (communication, compute, storage and network
services/applications paradigm), which contain a number of operational domains in all network segments (wire/wireless access,
core, edge, space or mixture of segments) , that may be accessed by a user from one or more locations.

Network Services Evolution
Infrastructure 4
Web
Multimedia APPs
Connectivityfor
Everything
Abundant
Bandwidth
everywhere
Critical Communication Services
Extreme and
Stringent QoS/
SLA
Secured/
Trusted
Network
Services
Industrial
Internet
Services
New Media /
Hologram
as a Services
New Telco Cloud
as a Service
Present Digital Network Services 2020-2030 Network Services 2030 AndBeyondNetworkServices
Network
Services
Evolution
Evolution
Evolution
`Tactile
Network
Services
• Delivery of stringent KPIs / SLAs per service (e.g. Gbps Tbps, less than 20 ms for round trip latency)
• Guarantees and monitor mission critical services;
• New Telco Precision Service Cloud, Industrial Internet Services, Hologram as a Service, New Secure Network Services
• Integration at Hyper Scale of elements for service de livery (i.e. network devices, network (virtual) functions, edge elements
and digital objects)
• Agility & Programmability (service functional change on demand)
• Anonymity and security support for all service operations

Cloud Network Slicing
Infrastructure
Network Slice Types Vs. Management Responsibilities
Provider
Internal
Slices
Provider
Slice As A
Service
External
Provider
Managed
Slices
External
Tenant
Managed
Slices
Tenant
Manages
Slices and
Services
Provider
Manages
Slices and
Tenant
Manages
Services
Provider
Manages
Slices and
Services
Provider
Manages
Slices as a
Service
Network Slice Types
Management
Responsibilities
5
Cloud Network Slice is a managed group of subsets of resources, network functions / network virtual functions at the
data, control, management/orchestration, and service planes at any given time. The behaviour of the network slice is
realized via network slice instances (i.e. activated network slices, dynamically and non-disruptively re-provisioned). A
network slice is programmable and has the ability to expose its capabilities.

NECOS Drivers
Driving issue
It is inefficient and expensive to build a separate infrastructure for
each service.
The NECOS project
Vision
Services delivered through Slicing will become the new norm
across resource types and administrative multi-domains cloud
networks as the way to deliver service specific KPIs.
Approach
Take the resource sharing paradigm to the next level of resource
sharing as a service. 6

Net App
Net App
NFs
Net App
Net App
L7 Apps
Network
Resources
NIM
SlicingSlicing
Application Services
Vertical
Use
Case i
Control & Management plane
Infrastructure
Business (Application & Service) plane
Slicing
Compute
Resources
VIM
SlicingSlicing
MonitoringMonitoringMonitoring
VIM-independent Slicing
[Mode 0]
[Infrastructure Slice aaS]
(“Bare-metal”)
VIM-dependent Slicing
[Mode 1] [Resource Slice aaS]
(R) Orchestration
Service-based Slicing
[Mode 3] [Service Slice aaS]
Network Service Orchestration
MANO-based Slicing
[Mode 2] [NFV aaS]
SlicingSlicing
SS
Vertical
SS
Service
iSS
Slicing Models & Approaches

NECOS Approach:
Lightweight Slice Defined Cloud (LSDC)
• LSDC a novel cloud network approach for Edge and Core that extends the virtualization to all the
resources in the involved networks and data centres, providing a uniform management with
advanced levels of orchestration.
• LSDC usage in
– abstracting, isolating, orchestrating and separating logical
behaviors from the underlying physical network& cloud resources
– creation of logically or physically isolated groups of network & service resources and (virtual)
network functions configurations
– helping adoption and integration of cloud computing in their large networks
– supporting edge devices with low computation and storage capacity
• LSDC cloud network - Differentiated Factors
– The Slice as a Service –- a new deployment model. Grouping of resources managed as a whole,
and that can accommodate service components, independent of other slices.
– Embedded methods for an optimal allocation of resources to slices in the cloud and networking
infrastructure, to respond to the dynamic changes of the various service demands.
– A Management and Orchestration approach making use of methods and artificial intelligence
techniques in order to tackle with the complexity of large-scale virtualized infrastructure
environments
– Making reality of the lightweight principle, in terms of small footprint components deployable
on large number of small network and cloud devices at the edges of the network 8

Project Use Cases and Requirements
9
Project addresses the limitations of current cloud computing
infrastructures to respond to the demand of new services, as presented
in two use-cases (UCs), instantiated in a set of scenarios that resulted in
50 elicited functional and non-functional requirements

NECOS Project High Level Architecture
10

NECOS Architecture and Interfaces
Defining a set of functional components and their interactions.

NECOS Information Model
12
Revisited specification of an information model that takes into account the views of the different
stakeholders. Used for designing Client-to-Cloud and Cloud-to-Cloud APIs
NECOS Slice Specification ModelNECOS Information Model

NECOS Slice-as-a-Service Model
Slice as a Service Interface Specification:
• the Slice Request Interface
• provides the mechanisms to initiate the instantiation of a Slice
• the Slice Marketplace Interface
• for interaction between marketplace actors to implement
mechanisms for the propagation of resource offerings between
resource domains.
• the Slice Instantiation Interface
• used to allocate resources for each single Slice Part of the Slice.
• the Slice Runtime Interface
• provides functionalities to dynamically modify the resource
allocation and to perform lifecycle operation on a Slice.

Slice / Features Selection/ Elasticity
workflows
14
The design ( i.e. internal functions) for the
components of the NECOS architecture and
workflows among these components
necessary to support the slice creation, slice
elasticity and slice decommission are
published in deliverable D6.2

Results & Achievements (I)
15
Objective 1: To develop and build a Lightweight Slice Defined Cloud (LSDC) platform
enabling computing, network, and storage elements in the cloud through the Slice as a
Service across federated clouds. This LSDC platform will have open APIs offering different
levels of service and resource abstractions and will enable the integration of applications with
slice resources for a faster and automated service provisioning in the networked cloud.
Results Published in Deliverable D2.2
• 2 Use Cases specification (Telco Cloud and MEC) with multiple scenarios (6); 50 functional & non-
functional requirements ( prioritization in D3.1)
• Identified 16 Key Performance Indicators, 10 Critical Success Factors, 4 Expected Differentiated Factors
• Business Analysis for LSDC
Results published in Deliverable D3.2
• An architecture of the LSDC was proposed
• Key components of this architecture are the Slicing orchestrator (Slice resource orchestrator, Service
orchestrator adapter, Slice specification processor, slice builder, slice database), Infrastructure and
Monitoring Abstraction (Resource and VM management, Resource and VM monitoring, Adaptors for
VIM/WIM control and monitoring interfaces), Resource Domains ( DC slice controller, WAN slice
controller), Resource Marketplace (Slice broker, Slice agent)
Results published in Deliverable D4.2
• Specification of an information model that takes into account the views of the different stakeholders
(Tenants, Slice providers and Resource providers).
• In support of the interaction among the LSDC platform and the slice tenants, the Resource Domains
and the Resource Marketplace we defined two categories of APIs, namely the Client to Cloud APIs and

Results & Achievements (II)
16
Objective 2: To develop the artefacts needed to make the LSDC a service
provisioning approach characterized by the integration of resources within the
collection of independent slices. A goal of the LSDC approach is to reduce the
complexity and timescale for service provisioning and deployment in federated
software defined DCs, and consequently reducing the OPEX for the infrastructure
owner.
Results Published in Deliverable 2.2
• Specification of multiple scenarios (5) : 5G Infrastructure vRAN Scenario, 5G Services Scenario, vCPE
Scenario, Touristic Scenario, Emergency Scenario. Scenarios provisioned on top of the NECOS
infrastructure, enabling the Telco Cloud and MEC environments.
• These scenarios are interpreted a set of interworking actors, namely the NECOS slice provider, the
Resource broker, the Datacenter infrastructure provider, the Network connectivity provider and the
Service provider / slice tenant
• Architectural specification of LSDC components, interfaces, and the associated systems
• Revised prioritization of functional & non-functional requirements
• Specification of a set of workflows that support service provisioning process: Slice Creation workflow,
the Slice Decommission workflow, the Slice Elasticity with Upgrade of Slice Resources workflow and the
Slice Elasticity with Downgrade of Slice Resources workflow.
• Design specification for all NECOS components (including Slice Provides components, Resource
Providers components, Marketplace components and Service orchestration component.

Results & Achievements (III)
17
Objective 3: To develop the service and resource orchestration
and management methods for the LSDC infrastructure of resources
within the network and at the edge of the network. This service
orchestration and management approach includes the automatic re-
allocation of resources and services across distributed and
geographically separated computing, data storage, and network
infrastructures within separate slices.
• Design of new supporting mechanisms for provisioning and orchestration, design of infrastructure and
monitoring abstraction component, design of slice resource orchestrator component, design of the
service orchestrator
• Realization of new interfaces for Slice-as-a-Service (as architected in D3.2)
• Design of the resource discovery framework – responsible the location of the appropriate resources
that compose a slice (i.e., slice components that correspond to service functions and service links,
according to the information model); of the resource marching model – responsible for mapping slice
request messages to the available providers’ resources; of a slice cost model – calculation of cost of a
slice based on its requirements and set of slice parts.

Results & Achievements (IV)
18
Objective 4: Demonstrate the full impact of
the NECOS solutions by way of the pilots
based on each of the use-cases
• Defining and realization of a set of reference implementations (5 demos) addressing specific subsets of
requirements of the NECOS platform.
• Each one of these demo reference implementations are tightly coupled to the concepts and
specifications mentioned in the achievement of the Objective 1 and the specified scenarios of D2.2. In
addition, we defined the minimal requirements of libraries, operating systems capabilities and the
network connectivity between interconnecting federated islands.
• 5 demos [Multi slice/tenant/service demo (MUSTS), Marketplace demo (MARK), Large-scale lightweight
Service Slices (ELSA), Machine-learning based orchestration of slices (MLO), Wireless Slicing Services
(WISE)] as a ground for demonstrating the use cases.
Open source solutions & demos
• 5 Prototype Demos will be partly presented in the workshop
• All demos are released as open source solutions https://gitlab.com/necos/demos
Net App
Net App
NFs
Net App
Net App
L7 Apps
Network
Resources
NIM
Slicing
Vertical
Use
Case i
Infrastructure
Slicing
Compute
Resources
VIM
Slicing
[Mode 0]
(“Bare-metal”)
(R) Orchestration
MANO-based Slicing
[Mode 2] [NFV aaS]
Slicing
S
Vertical
S
Service
iS

Dissemination and Standardization
impact
19
Dissemination Impact: 16 journal publications + 32 conference publications
Standardization Impact
• Y2 – 4 Contributions: “Network2030 Architecture Principles, including Slicing” - ITU-T Net2030; “Slice Provision Model” - IETF 104, “ E2E
Slicing Reference Framework and Information Model” - ETSI; “NECOS contributions to National Institute of Standards and Technology
(NIST) US Department of Commerce -NIST Cloud Federation Reference Architecture [Y1 - One completed and approved ETSI specification
“E2E Network Slicing Reference Framework and Information Model”; 16 contributions to IETF, and 1 contribution to ITU-T Net2030 FG]
Additional dissemination actions
• Y2 – 4 Workshops organized with NECOS participation: 1)Workshop on Theory, Technologies and Applications of Slicing for Softwarized
Infrastructures (WSlice 2019). Collocated with SBRC 2019, 06 May, Gramado-RS, Brazil; 2) IEEE/IFIP International Workshop on Hot Topics
in Network and Service Management: 5G and Sliced Networks, Distributed Things and Blockchains (HotNSM 2019), 12 April, 2019,
Washington, D.C., U.S.A. Collocated with the IFIP/IEEE International Symposium on Integrated Network Management; 3)2nd Workshop on
Advances in Slicing for Softwarized Infrastructures (S4SI). Co-hosted at the 5th IEEE International Conference on Network Softwarization
(NetSoft 2019), June 24-28, 2019, Orange Gardens, Paris, France; 4)2nd Workshop on Mobility Support in Slice-based Network Control for
Heterogeneous Environments (II MOBISLIE). Co-located with the V IEEE Conference on Network Function Virtualization and Software
Defined Networks, 12th November, Dallas, Texas, U.S.A. [Y1 - 5 Workshops organized with NECOS participation: 1st IEEE Workshop on
Segment Routing and Service Function Chaining (SR+SFC 2018) at IEEE CNSM 2018; ICT4I40 (1st edition of the Workshop on Information
and Communication Technologies for Industry 4.0) at IEEE ISCC 2018; Orchestration for Software Defined Infrastructures (04SDI), and 5G-
ready Network Applications Development and Orchestration over Network Slices with Mobility Support at IEEE NFV-SDN 2018; Workshop
on Advances in Slicing for Softwarized Infrastructures (S4SI) at NetSoft 2018 + 2nd edition at NetSoft 2019]
• 6 Keynotes organized and sponsored by NECOS in Y2: ACM MOBISLICE 2019, Futurecom 2019, ONF Connect 2019, IEEE SCLA 2019, IEEE
NetSoft 2019, IEEE ETNS 2019 ; 2 Keynotes organized and sponsored by NECOS in Y1 : IEEE NOMS 2018, IEEE SIU 2018)
• 5 Distinguished Expert Panels organized and sponsored by NECOS in y2: IEEE SCLA 2019, IEEE ICIN 19, IEEE NetSoft2019, IEEE WCNC
2019, ITU-T 5th Workshop Network 2030 (3 Distinguished Expert Panels organized and sponsored by NECOS in y1: IEEE CloudNet 2018,
IEEE NetSoft 2018, IEEE NOMS 2018
• 2 Slicing tutorials in Y2: IEEE SCLA 2019, IEEE NetSoft 2019 2 Slicing Tutorials organized and sponsored by NECOS in Y1 at IEEE
NetSoft2018, IEEE/IFIP CNSM 2018)

Slicing
Characteristics
20
The table shows
how the NECOS
architecture
addresses all of
the Slicing
Characteristics.
Few others
address cloud
slicing, network
slicing, or consider
end-to-end and
multi-domain
scenarios, or
embraced the
Slice-as-a-Service
paradigm.

Concluding Remarks
21
• Full partners technical cooperation
• 5 Demos with Open solutions released
• NECOS Key Value Proposition:
– Slice as a Service – From Concept to Realization in the context of Cloud
Networks
– Lightweight Slice Defined Cloud (LSDC) architecture and interfaces, new
integration of computing clouds with connectivity fabrics, Components
design, MEC & Telco Cloud UCs
Impact on SDOs (22) + Academic Publications (48) + Workshops (10) +
Distinguished Expert Panels (8) + Keynotes (8) + Tutorials (4) + Presentations.

Next Steps: 5G+ Networking
Service Adapted
Network Slices
Enabled by
Network Functions
Including NFV
Dedicated ICT Service
Network Slice
High-Precision Service
Network Slice
Mobility Network Slice
Light Weight Smart Network as a Service & APIs – Multi-domain Network
Operating System Facilities: Automation, Autonomicity, Network Abstraction & programmability,
Allocate (virtual) network resources/ slices, Maintain network state, Ensure network Reliability in a multi
domain environment
Management & Control
CORE
Smart Cloud &
Network Fabric
Enabled by
Programmability
EDGEMETRO
Execution
Environment
Node API
Node OS
RADIO
ACCESS
Execution
Environment
Node API
Node OS
FIXED
ACCESS
Execution
Environment
Node API
Node OS
Node API
Node OS
Execution
Environment
Execution
Environment
Node API
Node OS
Execution
Environment
Node API
Node OS
E2E Multi-Domain Orchestrator
E2E coordination, conflict resolution, multi-domain information exchange
Slice Cognitive and Autonomic

NECOS Demos
24
Demos:
• Multi slice/tenant/service demo (MUSTS)
• Marketplace demo (MARK)
• Large-scale lightweight Service Slices (ELSA)
• Machine-learning based orchestration of slices (MLO)
• Wireless Slicing Services (WISE)
• 2 demonstrations over the NECOS integrated test environment (i.e. MUSTS, MARK).
• 3 demonstrations over local lab test environment (i.e. ELSA, MLO, WISE)
• Each demonstration has an objective, list of NECOS components utilized, evaluation results, inputs
that will be provided by the tenant, and expected output. A tenant is responsible for the service
that will be running over the slice, in other words, this entity will play the role of a customer
requesting the slice in order to accommodate the deployment of a service.
• Demos specifications and evaluations are published in deliverable D6.3.
• Open-source demo solutions are published https://gitlab.com/necos/demos

Why slice-ready cloud network infrastructure
is needed?
• Vertical customers requesting services that lay outside the footprint of
their primary provider
• Interaction with other providers are needed but …
– How we can charge and bill for that service?
– How we can ensure SLAs among providers?
– How we can know about the capabilities of other providers for a
comprehensive e2e service provision?
• The current interconnection models is not aware of peer’s network
resources (i.e., load conditions, etc)
• All these environments are static, requiring long interactions for
setting up any inter-provider connection
• Automation for both the interconnection sessions and the service
deployment on top of that is needed to reach the goal of flexibility and
dynamicity
25

Concepts & Motivation
Concepts
• Infrastructure Slice - A set of infrastructure (network, cloud, datacentre) components/network
functions, infrastructure resources (i.e. managed connectivity, compute, storage resources) and
service functions that has attributes specifically designed to meet the needs of an industry vertical or a
service.
• Infrastructure Slicing - Management mechanisms that Infrastructure Slice Provider can use to allocate
dedicated infrastructure resources and service functions to Network Slice Tenant.
• Partition Types
– Physical separa on (e.g., dedicated backbones) → not cost eﬃcient
– A resource only partition is one of the components of a Network Slice → on its own does not
fully represent a Network Slice.
– Underlays / overlays, in the form of VPN as overlay solu on → not ﬂexible nor agile
– Slicing, through compute and network resource (including SF) alloca on → dedicated resources
per customer/service to ensure isolation on top of the same infrastructure
Roles
• Infrastructure Owner Owns the physical infrastructures (network/cloud/datacentre) and leases them
to operators. It becomes an ISP if it leases the infrastructure in a network slicing fashion.
• Infrastructure Slice Provider (ISP) – typically a telecommunication service provider, is the owner or
tenant of the infrastructures from which network slices can be created.
• Infrastructure Slice Tenant (IST) – is the user of specific network/cloud/datacentre slice, in which
customized services are hosted. Infrastructure slice tenants can make requests of the creation of new
infrastructure slice through a service model.
26

Project Use Cases and Requirements (2/2)
27
UC1 – Telco service provider is oriented towards the adoption of cloud computing in their large networks.
UC2 – Mobile Edge Computing (MEC) is targeting the use of edge clouds to support devices with low
computation and storage capacity.
5 Scenarios: 5G Infrastructure vRAN Scenario, 5G Services Scenario, vCPE Scenario, Touristic Scenario,
Emergency Scenario
4 Expected Differentiated Factors: DF1, DF2, DF3, DF4
10 Critical Success Factors: CSF1 – Isolation, CSF2 - Cost Reduction, CSF3 – Reliability, CSF4 - Flexibility,
CSF5 - Scalability, CSF26 - Elasticity, CSF7- Security, CSF8 - Slice Efficiency, CSF9 - Life-cycle Efficiency,
CSF10 - Simplicity
16 Key Performance Indicators: KPI1-Average elasticity response time, KPI2-Average end-to-end delay,
KPI3-Average service provisioning time, KPI4-Average slice provisioning time, KPI5-Average throughput,
KPI6-End-to-end slice availability, KPI7-Monitoring-data availability, KPI8 -Number of application users, KPI9-
Physical Server utilization, KPI10-Service demand prediction accuracy, KPI11-Service disruption index,
KPI12-Service QoE, KPI13-SLA fulfilment index, KPI14-Average Slice Decommission time, KPI15-Slice
isolation index, KPI16-Average Slice Provisioning time

Concluding Remarks (2)
28
As part of research on integrating Network computation with computing clouds the
following further challenges are envisaged:
• A Uniform Reference functional & non-functional Model for Large Scale Network Cloud Slicing.
• Uniform Slice Templates: Providing the design of slices to different scenarios.
• Efficient Slicing Service Mapping – creating an efficient service mapping model binding across network cloud slicing;
• Recursion, namely methods for slicing segmentation allowing a slicing hierarchy with parent–child relationships.
• Customized security mechanisms per slice - In any shared infrastructure.
• Network Slices Reliability - Maintaining the reliability of a network cloud slice instance, which is being terminated,
or after resource changes.
• Capability exposure for network cloud slicing (allowing openness); with APIs for slice specification and interaction.
• Programmability and monitoring control of Network Cloud Slices.
• Optimum and Light Weight Slice lifecycle management including creation, activation, deactivation, protection,
elasticity, extensibility, safety.
• Optimum Slice dimensioning.
• Autonomic slice management and operation, namely self-X for slices that will be supported as part of the slice
protocols.
• Very large scale slicing with efficient elasticity.
• Efficient and large scale slice stitching / composition

NECOS Problem Statement (1/2)
New Major Challenges in Cloud Networks
• Cloud Network is an integrated paradigm established to support the deployment of applications
and services without the need to rely on costly on premise infrastructures.
• Different services require different computing, connectivity and storage capabilities, with
different key performance indicators (KPIs) and different guaranties for QoS parameters.
• Best effort solutions for all services is not any more practical. It is also inefficient and expensive
to build and operate a separate cloud network infrastructure for each service.
→ Dynamic alloca on of resources from diﬀerent data centers (access, core, edge segments of the
infrastructure) and their interconnecting network systems to satisfy the needs of heterogeneous
service requests and all in an automated and efficient manner is a challenge that current cloud
computing platforms are not able to fulfil.
What is needed as Solutions
• New per service abstractions, isolating, orchestrating and separating logical behaviors from the
underlying physical resources.
• Efficient distributed multi-domain, end-to-end slicing as a service architecture and enablers,
where several data center providers would cooperate through the interconnection of different
partitions of their computation, connectivity and storage infrastructure for a service.
• Integrated management and orchestration of slice as a service (including Network and Cloud
resources)
29

NECOS Problem Statement (2/2)
Current State of the Research & Development
Fragmented landscape of Network & Cloud Slicing
→ risk of limited (and costly) inter-working solutions
• Not re-usable elements and systems, even at the conceptual level,
– Interoperability concerns
• Technology-specific solutions (e.g., 5G slices, Openstack-only)
• Heavy-weight solutions
• Standardization gaps
– multiple bodies with different scope: ITU-T, NGMN, ETSI, 3GPP, IETF, OMG,
OGF, TMF;
– no focused WG at any SDO,
– some overlaps and divergences.
30

Scope
Cross-domain management of slices in infrastructure
and service functions
31

Slicing Key Characteristics & Impact (I)
• Slicing has dynamic and non-disruptive re-provisioning
• Automation of operation
• Automated life-cycle management of slicing (Deploy, Change, Operate, Delete)
– Optimization resources (Auto-scaling/migration)
– Auto-healing
– Efficient Interplay between Management and Data Planes
• High-Scalability
– Separating to 1000~ slices (the order will vary depending on the use cases)
– Handling million ordered customers
• High-Reliability
– Redundant mechanisms
– High Level of Isolation
– Immediate fault detection
• Network Slicing with cross-domain by using open configuration model design
• Cost effective and prompt service/network deployment
– Utilization of virtualizing technologies (SDN and NFV)
– Harmonizing hardware and software appliances 32

Slicing Key Characteristics & Impact (II)
• Network/Cloud operators can exploit slicing for
– Enables other industrial companies to use networks as a part of their own services
(e.g. vertical players: connected car with high reliable network, online game with
ultra-low latency, video streaming with guaranteed bandwidth, etc.)
– Reducing significantly operations expenditures, allowing also programmability
necessary to enrich the offered tailored services.
– Means for network/cloud programmability to service providers and other market
players without changing the physical infrastructure.
• Slicing simplifies the provisioning of services, manageability of networks/clouds and
integration and operational challenges especially for supporting communication
services.
• Expecting realization of E2E slices and creation of new business model
33

NECOS Architecture – Full Design
Designing a set of functional components and their interactions. (Published in deliverable D5.2)

What do we mean by Cloud Network Slices?
Cloud Network Slice – A set of infrastructure (network, cloud, data center)
components/network functions, infrastructure resources (i.e., managed connectivity,
compute, storage resources) and service functions that have attributes specifically designed
to meet the needs of an industry vertical or a service.
A Network/Cloud Slice is a managed group of subsets of resources, network and service
functions at the data, control, management/orchestration, and service planes at any given
time. The behaviour of the slice is realized via infrastructure slice instances (i.e. activated
infrastructure slices, dynamically and non-disruptively re-provisioned).
– An infrastructure slice is programmable and has the ability to expose its capabilities.
– An end-to-end logical network/cloud running on a common underlying (physical or
virtual) infrastructure, mutually isolated, with independent control and management
that can be created on demand.
– A network slice may consist of cross-domain components from separate domains in
the same or different administrations, or components applicable to the access
network, transport network, core network, edge networks and clouds.
35

Types of Slices and Control Responsibilities
Additional viewpoints:
• From a business point of view, a slice includes a combination of all the relevant network & cloud
resources, functions, and assets required to fulfill a specific business case or service, including
management and DevOps processes.
• From the infrastructure point of view, infrastructure slice instances require the partitioning and
assignment of a set of resources that can be used in an isolated, disjunctive or non- disjunctive
manner for that slice.
• From the tenant point of view, infrastructure slice instance provides different capabilities,
specifically in terms of their management and control capabilities, and how much of them the
network service provider hands over to the slice tenant. As such there are two types of slices:
• (1) Internal slices, understood as the partitions used for internal services of the provider,
retaining full control and management of them.
• (2) External slices, being those partitions hosting customer services, appearing to the
customer as dedicated networks/clouds/datacentres.
• From the management plane point of view, infrastructure slices refers to the managed fully
functional dynamically created partitions of physical and/or virtual resources, network and service
functions that can act as an independent instance of a connectivity network and/or as a network
cloud. Infrastructure resources include connectivity, compute, and storage resources.
• From the date plane point of view, infrastructure slices refers to dynamically created partitions of
forwarding devices and servers with guarantees for isolation and security. 36

Analysis of the priorities
and relevance of requirements
37
• 50 elicited functional and non-functional requirements
• Analysis based on the Quality Function Deployment (QFD) methodology
developed by Y. Akao.
– For each scenario, we evaluated correlations between the identified requirements and
NECOS Critical Success Factors/NECOS Key Performance Indicators /NECOS Expected
Differentiated Factors (NECOS Characteristics).
– As such, we evaluated how each requirement is contributing to solve/enable each
Project Critical Success Factors/Project Performance Indicators/Project Expected
Differentiated Characteristics as seen from each scenario.
• Analysis Published in deliverable D3.1
0
50
100
150
200
250
300
RF.vRAN.1-Service Level Agreement
RF.vRAN.2-Accountability
RF.vRAN.3-On-demand slice provisioning
RF.vRAN.34-Isolation of slice provisioning
RN.vRAN.5 -Fairness
RN.vRAN.6 -Fault detection
RF.5G.1-Service Level Agreement
RF.5G.2-Accountability
RF.5G.3-On-demand slice provisioning
RF.5G.4-External control and management of…
RN.5G.5-Isolation of slice resources
RN.5G.6-Fairness
RN.5G.7-Fault detection
RF.vCPE.1-On-demand slice provisioning
RF.vCPE.2-Manageable slice
RF.vCPE.3-VIM-independence
RF.vCPE.4-Bare-metal slice
RF.vCPE.5-Lightweight virtualization
RF.vCPE.6-Elasticity
RF.vCPE.7-Zero touch service provisioning
RF.vCPE.8-Fault detection
RN.vCPE.9-Isolation of slice resources
RN.vCPE.10-SLA monitoring (QoS)
RN.vCPE.11-Low latency
RN.vCPE.12-High throughputRN.vCPE.13-High availabilityRF.Touristic(CD).1-Slice and slice-resource…RF.Touristic(CD).2-Automated Virtual Machine…
RF.Touristic(CD).3-Traffic load-balancingfor…
RF.Touristic(CD).4-Slice resource and service…
RF.Touristic(CD).5-Service planning
RN.Touristic(CD).6-Transparent end-user…
RN.Touristic(CD).7-Heterogeneity handling
RN.Touristic(CD).8-Elasticity
RN.Touristic(CD).9-Resource-efficiency
RN.Touristic(CD).10-Scalability
RF.Touristic(APP).1 Service function chain…
RF.Touristic(APP).2 Resource and user-demand…
RF.Touristic(APP).3 Resource offloading…
RF.Touristic(APP).4 Resource federation and…
RF.Touristic(APP).5 Scalability
RF.Touristic(APP).6 Efficient next-generation…
RF.Touristic(APP).7 Elasticity
RF.emergency.1 Dynamic slice management
RF.emergency.2 Dynamic service definition
RF.emergency.3 Timely slice management
RF.emergency.4 Orchestration
RF.emergency.5 High Reliability
RF.emergency.6 High Availability
RF.emergency.7 High Survivability
Average Score per Aggregated Requirement
NECOS Requirements- Importance in realising Expected Differentiated Characteristics
based on CombiningAll Scenarios
0
50
100
150
200
250
RN.vRAN.5 -Fairness
RF.5G.4-External control and management of the offered slices
RN.5G.6-Fairness
RN.vCPE.12-High throughputRN.vCPE.13-High availabilityRF.Touristic(CD).1-Slice and slice-resource managementRF.Touristic(CD).2-Automated Virtual Machine deployment
RF.Touristic(CD).3-Traffic load-balancingfor content delivery
RF.Touristic(CD).4-Slice resource and service monitoring
RN.Touristic(CD).6-Transparent end-user performance
RF.Touristic(APP).1 Service function chain orchestration
RF.Touristic(APP).2 Resource and user-demand prediction
capabilities
RF.Touristic(APP).3 Resource offloading between edge, core clouds
and cloud providers
RF.Touristic(APP).4 Resource federationand intelligent multi-
domain orchestration
RF.Touristic(APP).6 Efficient next-generation touristic application
performance
Average Score per NECOSRequirement
NECOS Requirements - Importance in realising Critical Success Factors based on Combining All Scenarios

Next Great Challenge: E2E Multi-Domain Slicing
Current wholesale and interconnection services and mechanisms
are not enough in the era of virtualization and programmability
• Vertical customers requesting services that lay outside the footprint of their
primary provider
– How to resolve this?
38

Slicing Key Characteristics & Impact
• A managed group of infrastructure resources, network functions and services (e.g.
Service Instance component, A Network Slice Instance component, Resources
component , Slice Capability exposure component).
• Concurrent deployment of multiple logical, self-contained and independent, shared or
partitioned networks on a common infrastructure platform.
• is a dedicated network part that is built on an infrastructure mainly composed of, but
not limited to, connectivity, storage, and computing.
• it is related to an operator that sees it as a complete network infrastructure and uses
part of the network resources to meet stringent resource requirements.
• Supports dynamic multi-service support, many/multi-tenancy and the integration
means for vertical market players.
• NS is programmable and has the ability to expose its capabilities. The behavior of the
network slice realized via network slice instance(s).
• Service customized Network Slices (enabled by NFV) + Smart Network Fabric for
coordinating/orchestration, control of network resource
• Guaranteeing service level for end to end across multiple (administrative) domains
• Flexible customizability
– automation as the way for simplifying the provisioning 39

Highlights - Summary
• All performed work is fully aligned to the DoW with
exception of WP5
• Change in the WP5 work focuss (i.e. platform
implementation design) for better alignment with the
rest of WPs with no changes to the WP5 objectives.
40
5
WP1 –> Project Management
WP7 –> Project Impact
WP3 –> Architecture &
components
WP4 –> Information Model
and I/Fs
WP2 –> UCs description& Requirements
WP5 –> Platform design &
implementation
WP6–> PoCs & Validation

Change in WP5 work focus
41
Change in the WP5 work focuss for better alignment with the rest of
WPs with no changes to the WP5 objectives
• WP5 was initially conceived to deploy a monitoring an abstraction layer (Task 5.1) and the intelligent orchestration
functionality for the LSDC platform (Task 5.2). Each of these tasks were also associated to the two deliverables of the project,
which adopted names associated to the above referred tasks. Therefore, D5.1 was initially conceived to report the design of
the monitoring framework and policies to be used in the context of the project showcases, and D5.2 to address the
orchestration and other functionalities.
• Nevertheless, at the early beginning of the project we realized two facts. The first fact was that starting any task in WP5
without clear reference architecture, as pursued in WP3, was very difficult and prone to make serious mistakes that would be
hard to revert. The second fact was that WP5 should be aligned with WP6 requirements because the main objective of WP6,
also stated in the DoA, is to showcase the ability of an integrated platform to serve both as a testbed of the integration of the
prototyping results and as a ground for demonstrating the NECOS use cases. The testing and demonstrations needed in WP6
are then requiring the implementation of specific functionality in WP5, likely a subset of functions of those specified in WP3.
In summary, to effectively and efficiently start WP5 we had to wait until WP3 and WP4 work reached a high mature level.
• Despite the limitations exposed above, we decided to start WP5 activity at the early beginning of the project to allow the
development teams to get acquainted with already existing development tools and development modules that could be used
in subsequent stages. By doing this, we could anticipate eventual challenges and make better decisions in the next steps.
• The project has been shaping these initial developments in alignment with the evolution of the NECOS architecture and its
testing and demonstration requirements. As a result of this approach we are able today to show the building blocks of what
we call “proofs of concept” units; in other words, several independent testing and demonstration modules. The platform
development strategy planned up to end of the project is based on two phases. The first phase, which has been reached at the
end of the project year one, is centred on the above-mentioned building blocks. The second phase will be reached at the end
of the project and will consist of the same blocks with augmented functionality or new ones if necessary.

Slicing Models & Approaches
Net App
Net App
NFs
Net App
Net App
L7 Apps
Network
Resources
NIM
SlicingSlicing
Vertical
Use
Case i
Infrastructure
Slicing
Compute
Resources
VIM
SlicingSlicing
[Mode 0]
(“Bare-metal”)
(R) Orchestration
MANO-based Slicing
[Mode 2] [NFV aaS]
SlicingSlicing
SS
Vertical
SS
Service
iSS

Requirements contributing most(✓) / least (✗) to
realising NECOS Critical Success Factors if a
Scenario is considered in isolation
43
Requirements contributing most ( ) / least ( ) to realising NECOS Critical Success Fanctors if a Scenario is considered in isolation
Scenario Scenario Scenario Scenario Scenario
Index Requirement ID & Name 5G Networks vCPE Scenario Touristic Scenario Emergency Scenario All Scenaros
1 RF.vRAN.1-Service Level Agreement
2 RF.vRAN.2-Accountability
3 RF.vRAN.3-On-demand slice provisioning
4 RF.vRAN.34-I solation of slice provisioning
5 RN.vRAN.5 -Fairness
6 RN.vRAN.6 -Fault detection
7 RF.5G.1-Service Level Agreement
8 RF.5G.2-Accountability
9 RF.5G.3-On-demand slice provisioning
10 RF.5G.4-External control and management of the offered slices
11 RN.5G.5-I solation of slice resources
12 RN.5G.6-Fairness
13 RN.5G.7-Fault detection
14 RF.vCPE.1-On-demand slice provisioning
15 RF.vCPE.2-Manageable slice
16 RF.vCPE.3-VI M-independence
17 RF.vCPE.4-Bare-metal slice
18 RF.vCPE.5-Lightweight virtualization
19 RF.vCPE.6-Elasticity
20 RF.vCPE.7-Zero touch service provisioning
21 RF.vCPE.8-Fault detection
22 RN.vCPE.9-I solation of slice resources
23 RN.vCPE.10-SLA monitoring (QoS)
24 RN.vCPE.11-Low latency
25 RN.vCPE.12-High throughput
26 RN.vCPE.13-High availability
27 RF.Touristic(CD).1-Slice and slice-resource management
28 RF.Touristic(CD).2-Automated Virtual Machine deployment
29 RF.Touristic(CD).3-Traffic load-balancing for content delivery
30 RF.Touristic(CD).4-Slice resource and service monitoring
31 RF.Touristic(CD).5-Service planning
32 RN.Touristic(CD).6-Transparent end-user performance
33 RN.Touristic(CD).7-Heterogeneity handling
34 RN.Touristic(CD).8-Elasticity
35 RN.Touristic(CD).9-Resource-efficiency
36 RN.Touristic(CD).10-Scalability
37 RF.Touristic(APP).1 Service function chain orchestration
38 RF.Touristic(APP).2 Resource and user-demand prediction capabilities
39 RF.Touristic(APP).3 Resource offloading between edge, core cloudsand cloud providers
40 RF.Touristic(APP).4 Resource federation and intelligent multi-domain orchestration
41 RF.Touristic(APP).5 Scalability
42 RF.Touristic(APP).6 Efficient next-generation touristic application performance
43 RF.Touristic(APP).7 Elasticity
44 RF.emergency.1 Dynamic slice management
45 RF.emergency.2 Dynamic service definition
46 RF.emergency.3 Timely slice management
47 RF.emergency.4 Orchestration
48 RF.emergency.5 High Reliability
49 RF.emergency.6 High Availability
50 RF.emergency.7 High Survivability

Requirements contributing most (✓) / least (✗)
to realising NECOS Key Performance Indicators if
a Scenario is considered in isolation
44
Requirements contributing most ( ) / least ( ) to realising NECOS Key Performance Indicators if a Scenario is considered in isolation
12 RN.5G.6-Fairness

Requirements contributing most (✓) / least (✗)
to realising NECOS Expected Differentiated
Factors if a Scenario is considered in isolation
45
Requirements contributing most ( ) / least ( ) to realising NECOS Expected Differentiated Fanctors if a Scenario is considered in isolation
12 RN.5G.6-Fairness

Requirements Priority: Requirements - Importance
in realising Expected Differentiated Characteristics
combining All Scenarios (II)
46
0
50
100
150
200
250
300
RF.vRAN.2-AccountabilityRF.vRAN.3-On-demand slice provisioning
RN.vRAN.5 -Fairness
RF.5G.4-External control and management of…
RN.5G.6-Fairness
RN.vCPE.12-High throughputRN.vCPE.13-High availabilityRF.Touristic(CD).1-Slice and slice-resource…RF.Touristic(CD).2-Automated Virtual Machine…
RF.Touristic(CD).3-Traffic load-balancingfor…
RF.Touristic(CD).4-Slice resource and service…
RN.Touristic(CD).6-Transparent end-user…
RF.Touristic(APP).1 Service function chain…
RF.Touristic(APP).2 Resource and user-demand…
RF.Touristic(APP).3 Resource offloading…
RF.Touristic(APP).4 Resource federation and…
RF.Touristic(APP).6 Efficient next-generation…
RF.emergency.7 High SurvivabilityAverage Score per Aggregated Requirement
NECOS Requirements - Importance in realising Expected DifferentiatedCharacteristics
based on Combining All Scenarios

Requirements Priority: Requirements - Importance
in realising Critical Success Factors based on
Combining All Scenarios
47
0
50
100
150
200
250
RN.vRAN.5 -Fairness
RF.5G.4-External control and management of the offered slices
RN.5G.6-Fairness
RN.vCPE.12-High throughputRN.vCPE.13-High availabilityRF.Touristic(CD).1-Slice and slice-resource managementRF.Touristic(CD).2-Automated Virtual Machine deployment
RF.Touristic(CD).3-Traffic load-balancingfor content delivery
RF.Touristic(CD).4-Slice resource and service monitoring
RN.Touristic(CD).6-Transparent end-user performance
RF.Touristic(APP).1 Service function chain orchestration
RF.Touristic(APP).2 Resource and user-demand prediction
capabilities
RF.Touristic(APP).3 Resource offloading between edge, core clouds
and cloud providers
RF.Touristic(APP).4 Resource federation and intelligentmulti-
domain orchestration
RF.Touristic(APP).6 Efficient next-generation touristicapplication
performance
Average Score per NECOS Requirement
NECOS Requirements- Importance in realising Critical Success Factors based on Combining All Scenarios

Concepts (VI)
Network Slice Representation
Forwarding Network
Element
Network Slice 1
Tenant A
Control
Infrastructure
Tenant B Control
Infrastructure
Network Slice 2
NF Network Function /
Virtual NF
NF1
NF2
NF4NF3
Network Service
Tenant A
NF1
NF2
NF4NF3
Network Service
Tenant B
Physical Network
EInfrastructure
48

Early Definitions of Network Slicing &
References(II)
ITU-T Slicing (2011) as defined in [ITU-T
Y.3011- http://www.itu.int/rec/T-REC-
Y.3001-201105-I] is the basic concept of the
connectivity and compute Softwarization.
Slicing allows logically isolated network
partitions (LINP) with a slice being
considered as a unit of programmable
resources such as network, computation
and storage.
Slice capabilities (2009) “Management and Service-aware Networking Architectures (MANA) for Future
Internet” – A. Galis et all - Invited paper IEEE 2009 Fourth International Conference on Communications and
Networking in China (ChinaCom09) 26-28 August 2009, Xi'an, China,
http://www.chinacom.org/2009/index.html
3 Slices Capabilities
– “Resource allocation to virtual infrastructures or slices of virtual infrastructure.”
– “Dynamic creation and management of virtual infrastructures/slices of virtual infrastructure across
diverse resources.”
– “Dynamic mapping and deployment of a service on a virtual infrastructure/slices of virtual
infrastructure.”
17 Orchestration capabilities
19 Self-functionality mechanisms
14 Self-functionality infrastructure capabilities
49

Early Definitions of Network Slicing &
References (III)
NGMN Slice capabilities (2016) - consist of 3 layers: 1) Service Instance Layer, 2) Network Slice Instance
Layer, and 3) Resource layer.
• The Service Instance Layer represents the services (end-user service or business services) which are to
be supported. Each service is represented by a Service Instance. Typically services can be provided by
the network operator or by 3rd parties.
• A Network Slice Instance provides the network characteristics which are required by a Service Instance
A Network Slice Instance may also be shared across multiple Service Instances provided by the network
operator.
• The Network Slice Instance may be composed by none, one or more Sub-network Instances, which may
be shared by another Network Slice Instance.
3GPP TR23.799 Study Item “Network Slicing’ 2016
ONF Recommendation TR-526 “Applying SDN architecture to Network Slicing” 2016
EU 5GPPP
• 30 Large Scale Research projects – all based on Network Slicing (https://5g-ppp.eu) (2015- 2018+)
• White Papers on 5G Architecture centered on network slicing (mark 1 - https://5g-ppp.eu/wp-
content/uploads/2014/02/5G-PPP-5G-Architecture-WP-July-2016.pdf) (2016) (mark 2 https://5g-
ppp.eu/wp-content/uploads/2018/01/5G-PPP-5G-Architecture-White-Paper-Jan-2018-v2.0.pdf)
(2018)
50

Infrastructure Slicing Value Chain
• Capability exposure: trough this utilization model, the providers can offer Application Programming
Interfaces (APIs) to the vertical business customers for granting the capability of managing their own
slices. Such management actions can include e.g. dimensioning, configuration, etc.
• Integration at customer premises: complementary network segments, in some cases pertaining to
the vertical business customer, become an integral part of the solution, requiring a truly convergent
network including the integration in existing business processes as defined by the vertical customer.
• Hosting applications: the provider offer the capability of hosting virtualized versions of network
functions or applications, including the activation of the necessary monitoring information for those
functions.
• Hosting on-demand 3rd parties /OTTs: empower partners (3rd parties / OTTs) to directly make
offers to the end customers augmenting operator network or other value creation capabilities.
51

NECOS Industrial Workshop Technical highlights by Prof. Alex Galis (University College London – UCL)

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to NECOS Industrial Workshop Technical highlights by Prof. Alex Galis (University College London – UCL)

Similar to NECOS Industrial Workshop Technical highlights by Prof. Alex Galis (University College London – UCL) (20)

More from Christian Esteve Rothenberg

More from Christian Esteve Rothenberg (20)

Recently uploaded

Recently uploaded (20)

NECOS Industrial Workshop Technical highlights by Prof. Alex Galis (University College London – UCL)