Future platform technologies and architectures

1
Future platform technologies and architectures
ITU Workshop on Service Delivery Platforms
(SDP) for Telecommunication Ecosystems from
todays realities to requirements and challenges
of the future(Geneva, Switzerland, 17 October
2011 )

• Roberto Minerva,
• Manager of IC Scenarios
• Telecom Italia Future Centre

2
Agenda

• A New Context for Telecomms
• Many paradoxes
• The Rising Importance of Data
• IoT and Data
• Personal Data Data Identity
• Deriving some requirements for Operators
  Platforms

R. Minerva, Telecom Italia / Future Centre
3
A New Context for Telecomms
4
Towards 2020 the future of networks

• Fiber reaches a large part of homes and
  enterprises, it guarantees
• Unlimited bandwidth
• Low cost per bit
• Everywhere connectivity
• flat rate also for mobile
• User is ABC (Always Best Connected)
• Transparent Connectivity (user always connected
  at the lower prices and best bandwidth)
• Strong Integration (at the terminal level)
  between mobile and fixed networks
• Cognitive Radio (terminals adapt to the available
  networks)
• Rapid change of Providers
• Dynamic Business Models
• Users connected to smaller, more efficient and
  cheaper cell
• Importance of Data
• Seamless access to personal data
• Capability to dynamically create, aggregate and
  update personal data
• New data type and data sets
• Integration between real and virtual related data
• User controlled access to data

• Services strongly related to users and their
  social relationships
• Identity Management and their association to
  Biometric techniques
• Different roles and identities per user
• Diminishing role of SIMs
• Fundamental Role of Terminals
• Abundance of Storage (1TB in the terminal, toward
  infinite in the network?)
• High processing Capabilities in the terminals
  (context controlled locally by the terminal)
• Capability to interoperate will be embedded in
  terminals and will be downloadable OTA
• Software embedded in the terminal makes the
  device more a personal service platform than a
  product. They will be personalized according to
  the specific user needs
• Pervasive Communication
• Ubiquitous connectivity
• Broad classes of communicating objects (smart
  object, beacons, smart materials, sensors,
  micromachinery, ...)
• Explosion of Augmented Reality and Internet of
  Things applications

Connectivity is a commodity Terminals are
flexible and personal service platforms Data
center larger and larger
5
Difference on Services
WebCos work on DATA
Telecoms work on BITS
6
Two Operators Problems

New Services New Revenue Streams SMART OPERATOR
Revenue
Cut Costs Optimize LEAN OPERATOR
Bit Pipe
t
2015
7
Two approaches
Lean Operator
Smart Operator

• Decoupling Services from the Network
• Global reach independent of the access network
• mashup of local and global resources (Net of
  Nets)
• Provide Global Services
• the long tail is much bigger
• because the marginal costs of a global solution
  are smaller (Skype the costs for building a
  global software are marginal)

• Networks will become highly dynamic Complex
  Systems made out of many heterogeneous networks,
  systems and intelligent endpoints
• Complexity will result in
• Costly infrastructure difficult to install,
  manage and integrate
• Lack of optimization of usage of resources
• Lack of knowledge of the network" as a whole
  and how it is globally and locally behaving in
  supporting customers requests

De-perimeterization of Services ? a Network
Operating System
0-Touch Networks
8
New Classes of Services

• Internet of Things
• Extended Reality
• Social Media
• e-Government
• Data (what s this?)
• Always Best Connected (ABC)
• Evolution of VPNs

Smart Operator
Lean Operator
9
Building a Data Path for Operators

• Exploit data related opportunities
• Move from Bits to Management of Information

10
The network itself is asource of Information
11
Mining the Operators Gold Mines

• Personal Data allow for Profiling the User
• CDR data
• Location Related info
• Usage of Internet Connectivity
• IPTV usage
• ...
• Statistical Data (related to more than one user)
• Usage of Network Resources
• Location Information and mobs movement
• ...
• Technologies
• Data Mining
• Reasoning
• Derive Data from Uncertainty
• Neutralization

http//www.seshat.ch/home/geom06.htm
12

• Internet of Things

13
Objects, Terminals, Networks as Personal Info
Producers
Intelligent Terminals
Smart Objects
NFC, Wifi, UWB, BT,
3G, 4G, other public nets,
Servers
Servers
Servers
Servers
Servers
R. Minerva, Telecom Italia / Future Centre
14
How Many Nodes, How Many Messages, How Much
Bandwidth ?
Issue low average traffic, but highly impulsive
traffic (e.g., SPIKES all objects wake up at 3
a.m. and send high priority messages)
Likely the communications between nodes will be
framed into B2B2C relationships Many
objects/nodes will come with communications
already paid for (I.e., embedded communications)
Equivalent Volume as the one generated by 20 M
users (3 min phone call per day)
Operators should try to be Aggregators by
providing real-time messaging engines (e.g.,
based on PubSub), by creating communities, and by
supporting the deployment of the infrastructure
(at home and in the public domain)
15
Where is the Value then ?
Knowledge
Information
Data

• Relationships

• Inference

• Aggregation
• Personalization

R. Minerva, Telecom Italia / Future Centre
16

• Some Issues with networked data

R. Minerva, Telecom Italia / Future Centre
17
A First Issue proper Management of Personal Data
Excerpt form Personal Data - The Emergence of a
New Asset Class (WEF)
Source Bain Company
A user-centric personal data eco-system
(WEForum)

• Volunteered data created and explicitly shared
  by individuals
• Observed data captured by recording the actions
  of individuals
• Inferred data based on analysis of volunteered
  or observed information

Personal Data should be properly managed
R. Minerva, Telecom Italia / Future Centre
18
But what is the Value of Personal Data?

• The value of the single datum depends on the
  usage context
• The kind of collected datum V(CreditCard) gt
  V(SocialSecurityNumber)
• Time reference V(CreditCard, today)gtgt
  V(CreditCard, 4years ago)
• The value of the Person in the usage context
• Importance of the person V(CreditCard, SB) gtgt
  V(CreditCard, RM)
• The linkage of the datum
• The connection of the datum and the person
• How many data can be connected to the datum (es.
  FoaF)
• Is the datum owner well connected (a Hub kind of
  person)V(a) gt V(b)
• How much data (derived from http//www.thoughtfarm
  er.com/blog/2009/04/27/intranet-roi/)
• Metcalfes Law n(n-1)/2
• Reeds Law 2n  n-1
• Briscoe, Odlyzko, and Tilly Law n log (n)
• Beckstrom Law V  SB SC

19
Identity of Things

• Each Resource is addressable
• Each resource is CONNECTED
• Connectivity must be guaranteed in a variety of
  environments
• Secure Links have to be guaranteed
• Each Resource can be associated to a User
  (Identity)
• Who owns these relations UserId - Location -
  ResourceId - data used/generated ?

• Things can collect user related actions and data
• Each Thing can be used for tracking Users
• Owners of Things can collect a lot of data

20
Dealing with Data Data Anonimization, Contracts
and Contextualization
R. Minerva, Telecom Italia / Future Centre
21
Nodes will connect each other in unpredictable
ways

• Many Objects scattered in the env (with local
  communication)
• Some Aggregators gathering and dispatching
  information
• Networks to cross

Aggregation 2
1
2
Aggregation 1
Increasing richness and complexity at the edge of
(Operators) networks
Aggregation 3
3
Node Aggregation at time t1
Network
1
Aggregation 2
3
Aggregation 1
2
http//muxware.net/sol_mesh.php
Node Aggregation at time t2
22
A Second Issue Complexity is moving to the edge

• Functionalities and data generation/consumption
  and storage in the edge
• Networks will comprise smart objects and
  processing, storage, communication resources
• Data and Info will be the user target and not
  mere connectivity
• Complexity coalesces at the edges (more nodes and
  more capabilities)
• While big networks are flattening and becoming
  simpler (less nodes)
• Local aggregation vs. average coverage
• Aggregation of intelligent points will be mainly
  unpredictable and even if high average
  connectivity capabilities will be provided, such
  a dynamic concentration will challenge the
  networks capabilities
• Competition for scarce resources for a limited
  period of time
• Terminals will have a key role
• Terminals as an integral part of the
  communication environment
• Mobility will be assumed for the majority of
  terminals
• Need to control new complex systems without human
  intervention
• The Operator's network needs to support and help
  in this new context

23
Dealing with Complexity at the edge
Self-Organization of Networks

• Management of complex and dynamic Network s of
  Networks will be critical
• No human intervention possible
• Competition on resources
• Require
• Self-organization
• Game theory techniques for highly distributed
  systems

http//innovation.gsa.gov/blogs/OCIO.nsf/dx/Manage
ment-Innovators-Bookshelf-Small-Pieces-Loosely-Joi
ned-A-Unified-Theory-of-the-Web-by-David-Weinberge
r-2002
R. Minerva, Telecom Italia / Future Centre
24
A Tussle Future Internet and Identity

• There is a need for an Identity Layer in the
  Future Internet
• For identify people
• For identify things
• To relate things, people and places
• For freeing people from Identity Providers
• People have the right to exist independently from
  a provider
• People are the owners of their identity (and
  names), homonymies should be managed in a far way
  (way just one Mario Rossi when there are plenty
  )
• Open framework
• Many options, and, in certain circumstances,
  support for Certification by a provider (a bank),
  the government, others
• Need for a lot of standardization and discussion
• Need to safeguard and protect the ownership of
  data

25
A Bit of Technology
26
Current Paradigms are not future proof
Control Layer
Network Intelligence (e.g., IMS) is a
hierarchical model based on the assumption that
control has to be exerted by a few specialized
control nodes
Resources
Physical Layer
Client Server model totally disregards the
network aspects and can easily lead to a tragedy
of commons (misuse of common networking resources)
Server
Client
network
R. Minerva, Telecom Italia / Future Centre
27
Ossification of Internet and Perspective for the
Future Internet

• Security. Currently it is tackled as an issue at
  the edge, while the network(s) could contribute
  to relieve some issues (e.g., DDOS)
• Mobility, the current Internet has not been
  designed for an optimal management of mobility,
  the Future Internet has to deal with a multitude
  of highly mobile objects (mobility built in)
• Network Identity, Users are not recognized and
  managed in the network, they are managed only at
  the edges (specific servers or applications)
• Integration of Applications and Transport/Control
  Layers. Currently there are not consolidated
  interfaces that allow for a better cooperation
  between the Network and the Apps. Many
  applications do not use resources properly (e.g.,
  p2p applications do retrieve data from far away
  hosts)
• Edges are becoming themselves Networks. There is
  the need to understand and manage the dynamics
  around Networks of Networks i.e., complex
  systems that impulsively request resources and
  use them while these resources have been
  designed to support statistically determined
  needs
• Focus on data and info and not on transport of
  bits

28
Enabling technologies

• autonomic capabilities and bio-inspired
  algorithms (e.g., gossiping, self-organization
  algorithms), to deal with complexity
• (self-organized) P2P overlays for clustering
  components, to guarantee scalability,
  reliability, and abstraction from underlying
  network
• resources virtualization, based on abstraction
  for coping with heterogeneity and on the
  definition of dynamic slices for multiple
  allocations
• programmable intelligent mechanisms, e.g.,
  based on auctions, game theory, etc., for
  optimized resource allocation and use
• cognitive cross layering, to allow the network to
  perceive conditions, decide and act autonomically
  to reach local/global/end-to-end goals in an
  optimal way, in cooperation with autonomic and
  self-organizing behavior of resources
• grid Computing to better integrate different
  resources (computing storage and network and to
  integrate new ones (sensors., actuators,
  micromachinery, ...)
• Information centric networking to better collect
  and use the needed wanted data information

29
New Control Patterns Publish Subscribe Model

• New Control Patterns that exceed the
  client-server and the network intelligence
  models
• Transaction oriented processing
• PubSub

Google code PubSUb Hubbub
R. Minerva, Telecom Italia / Future Centre
30
Information Centric Networking
http//www.xerox.com/innovation/news-stories/netwo
rking/enus.html
http//www.4ward-project.eu/index.php?id29
4ward project in this paradigm, the
communication abstraction presented to
applications is based on transfer of application
data objects instead of the end-to-end reliable
byte-stream used by the majority of applications
today.
Content Centric Networking A Self-Organizing
Network That Meets Information Needs What Is
It?A new approach to networking that enables
networks to self-organize and push relevant
content where needed.Content-centric networking
enables communication to happen anywhere,
anytime, and with any device - using any
available means.
R. Minerva, Telecom Italia / Future Centre
31
What Role and Architecture for Operators ?
32
A New Layering
TELECOM ITALIA
Resources Pooling from several Ownership Domains
R. Minerva, Telecom Italia / Future Centre
33
Operators Strategies
TELECOM ITALIA

• Service Provider
• Service Enabler
• BIT Carrier

The Network Architecture is influenced by the
Operators Strategy
34
Choose Your Role andYour Network
Bit Carrier

• Transport the data
• in the better way for the customer
• Helping Customer in retrieving INFORMATION
• Ease the work of finding data and information,
  linking them into something meaningful and manage
  them

Be good to transport lots of bits
Platform Provider
Build a Data oriented Platform (new networking
data handling) for others to use
Service Provider
Transform Data into Information and create plenty
of services by means of meaningful interfaces
35
A Bit CarrierNetwork
TELECOM ITALIA
Focus onTRANSPORT Few basicSERVICES Control
Platf.BASIC
R. Minerva, Telecom Italia / Future Centre
36
Service EnablersNetwork (a Network Operating
System)
TELECOM ITALIA
Platform APIs

• High Value Funct.by the NETWORK
• Information CentricNETWORK
• Beyond the CLIENT-SERVERmodel
• Customised levelsof API
• Control functions in the NETWORK
• New Communic. ModelsPUB-SUBSCRIBE
• Integration with other NETWORKS and TERMINALS

Abstraction
System APIs
Virtualized Resource Control Plane (Network OS)
New functions
Existing functions
Layer 3 / IP
Legacy
External Clouds
Switch
eNB
Layer 2 Carrier Transport
R. Minerva, Telecom Italia / Future Centre
37
How The Network Could Look Like
38
My 2 suggestions

• What T.I. is doing

• What should be done

• Working on alternative models for dealing with
  personal data
• Working on a Network OS with open APIs
• Working on zero-touch networks (e.g., AFI)
• Working on ICN (European projects)
• Working on a new view of Identity

• Future Internet will be dominated by Tussles
  looking at networks and architecture with more
  user flavor (involving more user associations ?)
• Propose an open framework for Identity and
  related issues (profiling, data ownership,)
• Push for a shift from bits to data and ICN
• A Great Initiative for User Controlled Network of
  Networks and Data ?

39
Thank you!
Roberto Minerva TORINO - Italy Phone 39 011
228 7027 Email roberto.minerva_at_telecomitalia.com
R. Minerva, Telecom Italia / Future Centre