JP Vasseur - Cisco Distinguished Engineer jpv@cisco

© 2007 Cisco Systems, Inc. All rights reserved. 1JP Vasseur - SENSORCOMM 2008 - France

“The Internet of Things – What if objects talked to each other ?”

Sensorcomm 2008France – August 2008

JP Vasseur - Cisco Distinguished Engineer [email protected]

mailto:[email protected]

© 2007 Cisco Systems, Inc. All rights reserved.JP Vasseur - SENSORCOMM 2008 - France 2

1 year after SENSORCOMM 2007 … Sensorcomm 2007 “Why IP for Sensor networks ?” We are now seeing major and fast progress on many fronts:

• “Why IP for Sensor Networks ?” => “The Internet of things”• New applications are emerging very quickly (e.g. Smart Grid)• IP has proven to be light enough to run on highly constrained devices,• Emergence of new LP L1/L2 technologies (LP Wifi, LP Bluetooth, PLC, …) advocating for an IP layered architecture,• Quick progress on the standardization front (IETF, ISA)• Formation of a new industrial alliance (IPSO)

On the dark side …• We do see proposals for protocol translation gateway (as planned !)• Research may want to have more papers on architectural issues


Internet

L2N

L2N

TrueMesh

Wireless HART

ISA SP100.11a

Xmesh

Znet

MintRoute

MultiHop LQI

CENS Route

Smartmesh

TinyAODV

Honeywell

Most promoters of non-IP solutions have understood that IP was a MUST: they call this “IP convergence”: A protocol translation gateway ! Or Tunneling …

So far … WAS (Wait And See) - The current Trend(IETF – 2007)

http://www.zigbee.org/about/profile.asp?memberID=1365&logo=/images/member_logos/dust.gif


New applications pretty much every day … but …

The number of proprietary solutions has literally exploded: Zigbee, Z-Wave, Xmesh, SmartMesh/TSMP, … at many layers (physical, MAC, L3) and most chip vendor claim to be compatible with their own standard

Many non-interoperable “solutions” addressing specific problems (“My application is specific” syndrome)

• Different Architectures, • Different Protocols

=> Deployments are limited in scope and scale,

More and more key players agree that IP offers flexibility (layered architecture), interoperability, performances and a wide set of proven tools and protocols: the standard of choice for the Internet of Things !


The number of applications for Sensor Networks is endless

New Knowledge

Improve Productivity

Healthcare

Agricultural

Energy Saving (I2E)

Predictive maintenance

Industrial Automation Health

Smart Home

Defense

High-Confidence Transport and assets tracking

Intelligent Building

Smart Cities

Smart Grid


Some applications have their own set of standards

Building Automation:• BACNet (ASHRAE/ANSI): defines services (who is, I am, Who has, I have used for device and object discovery). Services such as Read-Property and Write-Property are used for data sharing. Devices acting on objects (Analog Input, Analog Output, Analog Value, Binary Input, Binary Output, Binary Value, Multi-State Input, Multi-State Output, Calendar, Event-Enrollment, File, Notification-Class, Group, Loop, Program, Schedule, Command, and Device.). Support a number of L1/L2 layers, including ARCNET, Ethernet, BACNET/IP, P2P/RS232/Master-Slave Token Ring over RS-485 and Lonwork. • There is an over IP solution but IP is used as a layer 2 … Required BACNet routing. • We’re working with several key players to redefine BACNet 2 in a more IP friendly way.• LonWorks (ANSI/CEA-709.1-B): on twisted pair 78 KB/s, power line (3.6 and 5.4KB/s), FO, Infrared – Peer to Peer protocol for control command – 7 layer protocol stack.


Sensor Networks - Usually a constrained environment requiring adaptation• Energy consumption is a major issue (for non powered sensors/controllers),• Limited processing power (CPU, memory) (improved Hardware)• Prone to failures => very dynamic topologies,• When mobile => increase the dynamic nature of topology,• Data processing may be required on the node itself,• Sometimes deployed in harsh environments,• Potentially deployed at very large scale,• Must be self-managed (auto-discovery, self-organizing networks)

SensorsBattery/Power

Mote:MicrocontrollerStorageRadioClock

Enclosure


Can we make The Internet of Things a reality? YES ! With moderate effort …

• Do not try to find a solution to all potential problems: reduce the problem scope

• Adapt or reuse existing protocols … do not reinvent the wheel ! : DHCP-like, SNMP, ….

• Design new IP-based protocols when needed:Example ? Routing … (see next slides)

• Preserve the fundamental openness of IP

• IP is ubiquitous and Sensors are everywhere … Good match.


Specify new IP protocols when needed

Routing in Sensor Networks is a MUST for energy saving (short distances => less energy to transmit) *and* to route around obstacles (including poor quality links),

Highly constrained devices Harsh & dynamic environments: (variable link qualities, link/nodes fail at

a rate significantly higher than within the Internet) Small MTU (high error rate, limited buffer/bw) Constraint routing is a MUST: take into account link *and* nodes

properties and constraints (also unusual)

Deep power management: WSN in sleep mode most of the time

Let’s face a reality: routing in Sensor Networks has unique requirements:


Highly heterogeneous capabilities Structured traffic patterns: P2MP, MP2P but also more and

more P2P

Multi-path and asymmetrical load balancing

Data aware routing: data aggregation along a dynamically computed path to a sink.

Self-Managed !!

Specify new IP protocols when needed (Cont)


Routing for Smart Objects

Current InternetNodes are routers

IGP with typically few hundreds of nodes,Links and nodes are stable,Nodes constraints or link bandwidth are typically non issues,Routing is not application-aware (MTR is a vanilla version of it)

Sensor NetworksNodes are sensor/actuators&routersAn order of magnitude larger in term of number of nodes,Links are highly unstable and Nodes die much more often,Nodes/Links are highly constrainedApplication-aware routing, in-Band processing is a MUST


The “Mesh Under” versus “Route Over” Debate or again “Why do we need IP” ?

• Many times people have argued in favor of a layer-2 approach for Sensor Nets, at best with IP reachability,

• Sensor Networks are made of a variety of links: wired and wireless,

• Even for WSN, there won’t be a single “winner”: IEEE 802.15.4, LP Wifi, Wibree, …

IP routing is a must for Sensor Networks


Combine “Mesh Under” and Route Over”: again ?

IP Routing over 802.11s, 802.16J, 802.15.5, Zigbee

• IP layer with no visibility on the layer 2 path characteristic

• Makes “optimal” routing very difficult

• Layer 2 path (IP links) change because of layer 2 rerouting (failure or reoptimization) lead to IP kink metric changes. How is this updated ?

• Remember IP over ATM

• There is still a need for an abstraction layer model but for Point to Point layer 2 links


Combine “Mesh Under” and Route Over”

Another major challenge: multi-layer recovery

• Require a multi-layer recovery approach

• Current models are timer-based:Needs to be conservative and most of the time bottom-upIncreased recovery time for failures non recoverable at layer 2

• Inter-layer collaborative approaches have been studied (e.g. IP over Optical) => definitively too complex for current Sensor Hardware


IETF Update

• Reuse whenever possible, Invent where needed

GEN OAM INT RTGAPS RAI TSVSEC

Reuse Existing WG dealing with LLNs

6lowpan ROLL


IPv6 over Low power WPAN (6LoWPAN) Additional information is available at tools.ietf.org/wg/6lowpan

Chair(s):Carsten Bormann <[email protected]> Geoffrey Mulligan <[email protected]>

Internet Area Director(s):Jari Arkko <[email protected]> Mark Townsley <[email protected]>

Internet Area Advisor:Mark Townsley <[email protected]>

Secretary(ies):Christian Schumacher <[email protected]>

Mailing Lists:General Discussion: [email protected] Subscribe: [email protected] Body: subscribeArchive: https://www1.ietf.org/mailman/listinfo/6lowpan


WG Status The Problem Statement document made RFC 4919 The Format Document is RFC 4944 Just re-chartered:

1. Produce "6LoWPAN Bootstrapping and 6LoWPAN IPv6 ND Optimizations”

2. Produce "6LoWPAN Improved Header Compression" to describe mechanisms to allow enhancements to the 6LoWPAN headers.

3. Produce "6LoWPAN Architecture" to describe the design and implementation of 6LoWPAN networks.

4. Produce "Use Cases for 6LoWPAN" to define, for a small set of applications with sufficiently unique requirements,

5. Produce "6LoWPAN Security Analysis"

Off-charter discussion on fragmentation ,flow control, …


Routing Over Low power and Lossy Link (ROLL) WG

Working Group Formed in Jan 2008http://www.ietf.org/html.charters/roll-charter.html Co-chairs: JP Vasseur (Cisco), David Culler (Arch Rock)

Work ItemsRouting Requirements ID for Connected HomeRouting Requirements ID for Industrial applicationsRouting Requirements ID for Urban networksRouting Requirements ID for Building AutomationSurvey on existing routing protocol applicabilityRouting metrics for LLNsRouting for LLNs Architecture document

Active work with a good variety of participants Already four WG documents as of May 2008.

We did limit the scope !


Slides about the protocol survey ID (IETF-72)

© 2007 Cisco Systems, Inc. All rights reserved.JP Vasseur - SENSORCOMM 2008 - France 24IETF 72 – July 2008 - Dublin

New Routing Metrics Motivation of the document

Unique characteristics of LLNsTypical routing metrics such as hop counts or link metrics are not sufficient for LLNs

A new set of required link and node metrics suitable to LLNs needs to be specified

ROLL WG itemNov 2008 Submit Routing metrics for LLNs document to the IESG to be considered as a Proposed Standard.

Classification of routing metricsLink versus Node metricsQualitative versus quantitativeDynamic or static


IETF 72 – July 2008 - Dublin

New Routing Metrics Routing metrics for LLNs is a critical topic

Need to be cautious !!! May be tempting to define a plethora of metrics … but not always implementable and usable in a deployed network

Use of dynamic metrics have been studied and experimented in the past (ARPANET: first average delays, revised metrics)

Dynamic metrics => Use of energy … The challenge is not to define metrics but to compute these

metrics.

This first revision lists potential candidates

© 2007 Cisco Systems, Inc. All rights reserved.JP Vasseur - SENSORCOMM 2008 - France 26IETF 72 – July 2008 - Dublin

A first list of Node and Link Routing Attributes Node metric

Computational resourcesResidual Energy (dynamic)Current workload (dynamic)Node latencyData Aggregation attributeNode degree Dynamicity Node reliability

Link metrics BandwidthReliability (Quality) Propagation delaySet of costs (missing from the ID)


ROLL: Next Steps

Try to Last Call all application specific routing requirements documents by September 2008

Call for a ROLL Interim WG meeting in October (TBC)

Draw a consensus


What about ZigBee, Z-Wave, and other proprietary protocols ?


One physical layer will not fit all Different requirements (range, power, bandwidth, frequency band, media, security)

Lead to different physical layers: Power line, 802.11, 802.11LP, 802.15.4, 802.3, 802.15.4a

IP is the common

abstraction layer


One solution: IP IP is independent of the physical layer.

Does IP work on highly constrained devices (meaning small and with limited memory) such as a 16-bit micro-controller with tens of Kbytes of RAM and possibly battery operated?

Absolutely: this has been very successfully demonstrated and there ARE several deployments!

The suggested IP stack only needs about 4K RAM

Will there be new IP protocols?MANY of the existing IP protocols can be used with no cost such as SNMP and PingNew IP protocols will be implemented only when and where needed.

IPeverywhere


Work at the IETF

A lot of us are very active at the IETF:In the 6lowpan Working Group (IPv6 over low power radio)A new routing WG has been established - ROLL – which has been very active and has Cisco representative as co-chair

Also at ISA (defining a standard for machine to machine communication in industrial environments) - ISA just adopted IP for that implementation

In both those areas there is a strong momentum and new companies joining the effort!


Why do we need a new alliance?

The IETF defines protocols – Does no Marketing

Regular requests for white papers, lists of companies supporting IP for Smart Objects, and other marketing support

We care about the end-user


Proposal Form a new Alliance promoting IP for Smart

Objects (also known as “Sensor Networks” and “The Internet of Things”)


Objectives and Areas to Avoid Objectives of the Alliance Promote the use of IP in Smart Objects by

publishing white papers, case studies, issuing technology press releases, providing updates on standards progress and other supporting marketing activities

Organize focused interoperability testing events

Areas to Avoid - the Alliance will NOT work on protocol specifications, algorithms, etc. – those activities will be done at the IETF !


Conclusion Sensor networks have a tremendous number of

opportunities but it is time to react and change the current “trend” (myriad of proprietary worlds),

The “WAS” approach (unavoidably leading to translation protocol gateways, complex tunneling, cross layer adaptations) is now strongly questioned.

We (hopefully) learnt from the past: open-based standard is KEY. IP is obvious the protocol,

Strong momentum around IP is being built: fast progress at the IETF (6lowpan and ROLL), formation of a new industrial alliance promoting IP (IPSO),

Participation of the Research community is key.

JP Vasseur - Cisco Distinguished Engineer jpv@cisco

Documents

Transcript of JP Vasseur - Cisco Distinguished Engineer jpv@cisco