Community Networking: Getting Peer to Peer out of Prison

1
Community Networking Getting Peer to Peer out
of Prison

• Marie-José Montpetit Eng. Ph.D.
• Motorola HNM Technology Office

2
P2P?

• Peer-to-peer (P2P) is a type of transient
  Internet network that allows a group of users
  with the same networking program to connect
  directly with each other to provide a certain
  service (originally file access but now voice,
  remote execution etc.) (from whatis.com)
• P2P is a loaded word that really defines how a
  community of users shares resources
• Move from thiefs to collaborators
• P2P (and collaborating edge devices) are here
  already
• Bluetooth devices and connectivity for file
  transfers between phones, PCs and even Lego
  robots
• DLNA and UPnP home networks
• Network resources pooled in meshed WIFI networks
• SETI
• Etc.
• The challenge is to demonstrate the P2P
  advantages
• across the value chain and the OSI stack but
  advantages to who?

3
Why Community Based?

• Just need to re-interpret (and reinvent) P2P?
• Move the network functionality to the edge
• Control plane (p2p signaling for discovery)
• Management plane (autonomics aspects)
• Data plane (p2p routing)
• Exchange deploying more expensive uplink
  bandwidth for shared community resources
• The need for sharing user generated multimedia
  (YouTube etc.) will not go away
• Uplink bandwidth is still expensive but local
  bandwidth (WIFI for example is cheaper)
• Less people pay but they pay more
• The community manages the local costs (SON
  heritage?)
• Use hybrids
• Uses some servers to offer connectivity to
  outside services (PVRs for example) while serving
  the community
• Maybe essential if content is commercial
• Defines multisided applications without the
  center

4
Why Community Based?

• What is a community?
• Sneaker net
• Home net
• WPAN, WLAN, WWAN
• Neighbors and family
• Global social network
• The ecosystem
• Gateways
• Handsets
• PC
• STB
• Servers
• Sensor networks
• Appliance networks
• and even webcams!

5
Why Community Based?

• Need to get away from the Lord of the boxes
  approach
• No need to have the device that rules them all
• Need to harness the power of
• Home devices and home networks connectivity
• Power users
• Focus on
• Autonomics
• Social networking
• Communities of interest
• Scalability of the architecture
• But (still) cost can be an issue
• And the absence of common platform introduces
  complexities
• Use case
• Laptops for CPU intensive applications
• DVR for storage (local and non local)
• Cell phones for voice
• Web pads for display
• Sensors for real time information
• Etc.

6
A solution Communities at Different Layers

• Physical layer the resources and the devices
• Network layer the connectivity
• Service layer the social fabric

7
Peering at Different Layers

• Physical layer
• Manage the installed base of broadband
  connections to the neighborhood
• Survive the uplink glut
• Reduce the number of high speed lines needed to
  serve a neighborhood
• Use super peers for local distribution
• Reduce the number of streams serving the
  community
• Use peer based storage/DVR
• Watch your neighbors TV as an alternative to
  server based channel change
• It is well known that not all of the 500
  channels are watched at any instant
• Encourage the use of home networks, local CPU and
  shared resources
• Reductions in OPEX/CAPEX passed down to the
  users?

8
Peering at Different Layers

• Network layer
• Advanced policy management
• Favor the community
• Device and service discovery 2.0 not just what
  they are, who they are and how to access them
• Heterogeneous peers
• Capability discovery
• Determine if this device is a renderer, a transit
  point and or a storage point, as well as
  determining if is has a link to the non P2P
  network (for AAA for example) and can become a
  super node and a trust authority
• Local routing and load balancing
• Scalability
• From sneaker net to SETI
• Most likely tens to thousands
• Propagation of discovery information
• QoS
• Some operators have a less than best effort
  category for P2P
• Video quality can suffer
• Cases to consider
• Pre-recorded known traffic profile
• Live content
• Unknown traffic profile especially for video

9
Peering at Different Layers

• Service layer
• Use your social network as a virtual operator
• You and your friend define what the offering is
  and transfer it to the nearest peer
• Based on common likes
• Based on instanteneous behavior
• Multi-network and multi-operators
• Define common middleware across platforms
• A challenge your community will most likely not
  have a uniform installed base of devices
• For the moment all applications are developed in
  silos
• Features
• Autonomics
• Security and privacy and of course DRM when
  applicable
• Local media switching (no more channel change,
  content change)

10
Business Aspects

• What could make operators adopt a community
  approach
• Uplink glut
• Sticky-ness of their customers (themselves
  becoming over the top)
• New provisioning and customer management
• QOS/policy management
• New method to distribute content?
• Put content in front of more customers event
  outside their networks
• PVR are now secondary content and video servers
• But
• Are new P2P entrants eroding the business of
  the incumbents?
• V.g. Joost, VEOH, Neocast vs. Comcast, Verizon,
  FT, BT etc.

11
Business Aspects
12
Business Aspects

• Recent work from Chintan Vaishnav of Sloan/CSAIL
  on using game theory to evaluate the impact of
  new entrants on incumbents business
• Players
• Incumbents, entrants, content providers (for
  video) and end users
• Early conclusion
• P2P Incumbent shifts the focus of the service
  from the network and bandwidth to ancillary
  services and quality of the service offering
• Move from a winner take all to a more
  diversified marketplace

13
Business Aspects (sample result)
14
Conclusion

• Where do we go now?
• Evolve the network offering to the community
• Video is key better use of the scarce uplink
  resources and untapped local resources (PVRs,
  WIFI etc.)
• Offer some new paradigms for management of
  communities and discovery of peers and their
  capabilities (not all the peers are equal)
• Making sure the devices that are part of the
  community are used appropriately
• Create value and revenue
• Is the community the key to the next generation
  Internet and the move to the Mobile Internet?