Opportunities and challenges in embedded systems

Opportunities and challenges in embedded systems-

logoTBD

Gerrit MullerUniversity of South-Eastern Norway-NISE

Hasbergsvei 36 P.O. Box 235, NO-3603 Kongsberg Norway

[email protected]

Abstract

The technological advances in processing, communication, storage, actuatingand sensing enables a large amount of applications of embedded systems. Thechallenges of today to realize these opportunities are discussed, addressing sixmain issues: market dynamics, interoperability, reliability, power, security, andcreativity.The capabilities of the Embedded Systems Institute are discussed briefly.

DistributionThis article or presentation is written as part of the Gaudí project. The Gaudí project philosophy is to improveby obtaining frequent feedback. Frequent feedback is pursued by an open creation process. This document ispublished as intermediate or nearly mature version to get feedback. Further distribution is allowed as long as thedocument remains complete and unchanged.

All Gaudí documents are available at:http://www.gaudisite.nl/

version: 1.0 status: concept September 1, 2020

1 Introduction

The field of embedded systems design is very broad, ranging from small systems,such as chips, unto large systems such as MR scanners or wafer steppers. Thedesign of these systems can be characterized by the combination technology intensive,to interface and interact with the physical world, and software intensive, to createthe desired functionality and integration. A lot of skills are needed to create thesesystems where all the emerging qualities of the system match with the needs of theuser of the system.

software

smart(?)

sensors

micro-sizeactuators

displays

processing

power

GigaOps/s

communication

bandwidth

GigaBits/s

storage

GigaBytes

infinite(?)

embedded

opportunities

Figure 1: The fast improvement of many fundamental technologies creates Gigaembedded opportunities

In this article we identify the challenges in Embedded Systems Design, basedon the trends in the world both at the customer side and at the technology side. Thestarting point of the analysis is the observation that the tremendous technologicaldevelopments create many opportunities, see Figure 1.

creativity market dynamics

discover latent needs

enable emergence

where is the business

globalization

hype waves

Moore's law

power consumptionweight, cost, performance

reliability complexity

heterogeneity

#engineers involved

interoperabilityemerging behavior, future vs legacy

heterogeneous vendors

securityprivacy, DRM

versus usability

Figure 2: Hit list of challenges

The world of high tech product development faces a number of challenges,summarized in figure 2. Section 2 discusses the market dynamics. Section 3

Gerrit MullerOpportunities and challenges in embedded systemsSeptember 1, 2020 version: 1.0

University of South-Eastern Norway-NISE

page: 1

expands the interoperability issue. The reliability is discussed in Section 4. Powerconsumption is elaborated in Section 5. Section 6 addresses the tensions betweenthe many security related viewpoints. The analysis is finished in Section 7 bydiscussing the creativity as limiting factor in product creation.

This article reuses previous material from [2] and [3].

2 The Dynamic Market

In this section two examples of a dynamic market are given. The first exampleis the semiconductor market, discussed in Subsection 2.1. The second example isfrom the medical domain, described in Subsection 2.2.

2.1 The Dynamics in the Semiconductor Market

Philips Semiconductors (PS) plays a part in a longer value chain as depicted infigure 3. Typical the components of PS, such as single chip TV’s, are used bysystem integrators, which build CE appliances, such as televisions. These appli-ances can be distributed via retail channels or via service providers to end consumers.

ProvidersUPC

Canal+ AOL

AT&T

RetailersFry's

Dixon

Consumers Boonstra

PeperKok

Chirac

Blair

PietersenSmith

Jones

Jansen

Muller

Kleisterlee

Clinton

v.d. Spijker

Meulengraafder Kinderen

Reinders

Bush

Rooyakkers

de Vries

Koch

d'Oliviera

van Oranje

Obbink

v.d. HamerCharite

Cruijf

Neeskensvan Hanegem

Goedkoop

Sharon

El Khatabi

de Gruijter

Heijn

Schijvens

Waterreus

Leonardo

van Bommel

Nistelrooij

Gandhi

Pinochet

Bakker

v.d. Meulen

Hoessein

Schroder

Schweitzer

Peters

Gore

System IntegratorsSony Philips CE-DN

Loewe

NokiaPhilips CE-TV

Philips CE-PCC

Component and

Platform SuppliersPhilips Semiconductors

Philips Components ST

TI

Samsung

Microsoft

Intel

Liberate

Micron

LG

It's

Prodi

Figure 3: Value chain

One of the major trends in this industry is the magic buzzword convergence.



page: 2

Three more or less independent worlds of computers, consumer electronics andtelecom are merging, see figure 4; functions from one domain can also be done inthe other domain.

Telecom

Consumer

Computer

Figure 4: Convergence

The name convergence and the visualization in figure 4 suggest a more uniformset of products, a simplification. However the opposite is happening. The conver-gence enables integration of functions, which were separate so far for technicalreasons. The technical capabilities have increased to a level, that required function-ality, performance, form factor and environment together determine the productsto be made. Figure 5 shows at the left hand side many of today’s appliances, in themiddle many form factors are shown and the right hand side shows some environ-ments.

mp3

dvd

set top box

flat display

pen

speech

cable

modem

firewall

Ambient Intelligence

living room

car

car navigation

pda

surveillance

camera

camera

GSM phone

computerCommunicator

television

games

sailboat

audio

microset

headphone

garment

watch

Figure 5: Integration and Diversity

Note that making all kind of combination products, with many different formfactors for different environments and different price performance points creates avery large diversity of products!



page: 3

AOL

Amazon.com

source: BigChart.com

dd march 19, 2001

1997 1998 1999 2000 2001

Figure 6: Uncertainty (Dot.Com effect)

Another market factor to take into account is the uncertainty of all players inthe value chain. One of the symptoms of this uncertainty is the strong fluctuationof the stock prices, see figure 6.

time to

marketeffort

10 100

digital TV

GSM

application

infrastructure

1 10

100

manyearmonths

volume

units

106

103

personalized

(skins, themes)

TV

1000

1

GSM

Figure 7: System Integrator Problem Space - Business

From business point of view the products and/or markets of the system integratorscan be characterized by time to market, volume, effort to create. In these 3 dimen-sions a huge dynamic range need to be covered. Infrastructure (for instance thelast mile to the home) takes a large amount of time to change, due to economicalconstraints, while new applications and functions need to be introduced quickly (tofollow the fashion or to respond to a new killer application from the competitor).The volume is preferably large from manufacturing point of view (economy ofscale), while the consumer wants to personalize, to express his identity or community



page: 4

(which means small scale). As mentioned before the effort to create is increasingexponentially, which means that the effort is changing order of magnitudes overdecades. Figure 8 summarizes these characteristics.

106

109

1012

performance power

10-3

1

home

server

digital TVhome

server

digital TV

GSM

GSM

storage

106

109

1012

digital TV

home

server

GSM

Operations/s Watt Byte

103

Figure 8: System Integrator Problem Space - Technology

Main technology concerns of the system integrators are performance, powerand storage. Again a huge dynamic range need to be covered in these dimensions.Video based applications have much higher processing demands than GSM speechaudio. While for power portable appliances like a GSM have severe constraintsand should use orders of magnitude less power than TV’s or set top boxes. Theamount of storage is again highly function dependent, for instance a home serverwhich must be able to store many hours of video needs a huge amount of storage,while the address book of a GSM phone is very limited in its storage needs. Thetechnology parameters and dynamic range are visualized in figure 8.

Combining the figures in one picture enables the visualization of a systemprofile. Figure 9 shows the profiles for a digital TV and for a GSM cell phone.

2.2 Dynamics in the Medical Market

The market is changing continuously and the frequency of changes is high. Figure 10shows several forces operating on the PMS market.

Companies from other domains are changing the playing field. Especially gentechnology (early risk indication, preventive examination and therapy?), pharma-ceutics (focused and localized application of medicines) will change the clinicalfield quite a lot.

Information Technology suppliers and service providers, integrate all types ofinformation and applications in networks of systems. This integration enables newapplications and also influences the conventional system boundaries and functionallocations.



page: 5

Problem space

106

109

1012

home

server

digital TV

GSM

Operations/s

10-3

1

home

server

digital TV

GSM

103

106

109

1012

digital TV

home

server

GSM

Byte

10

application

infrastructure

1

100

monthsunits

106

103

TV

1

GSM

personalized

(skins, themes)

100

digital TV

GSM

10

manyear

1000

pe

rform

an

ce

po

we

r

sto

rag

e

time

to

ma

rke

t

vo

lum

e

effo

rt

Watt

Figure 9: System profile

Government have a strong influence on the health care systems. Current trendsare towards cost reduction and cost control in attempt to keep health care provisionsaffordable for the entire population. At the same time increased privacy awarenessand concerns result in new and more regulations, with far stretching consequencesfor most clinical systems. The current threat of terrorism will also influence systemrequirements.

A global trend is an aging population, which has a considerable impact onthe consumption of clinical services. Also new diseases appear, which createsunforeseen needs.

gen technology

pharmaceutics

information integration

new materials

miniaturization

new sensors

wireless

cost reduction

cost control

privacy regulation

anti-terrorism regulation

aging population

new diseases

ne

w

tech

no

log

y

other

domains

go

ve

rnm

en

t

global

trendsPMS

Figure 10: Market dynamics in medical



page: 6

The rapid advance in many technology areas (for example new materials, minia-turization, new sensors, and wireless) opens up many new possibilities.

3 Interoperability

Systems get more and more directly connected. In the past humans linked thedifferent systems together, applying intelligence and adaptations were needed. Inthe electronic age the adaptation must take place in the systems themselves.

world

hospital

radiology department

MR scanner

acquisitionrecon-

structiondisplay

host

workstation storage printer

workstation archive

PC beamerRIS

other

examination

rooms

HIS LIS

other clinical

departments

PC security admin

clinicians at

home

patients at

home

patients

away

clinicians

away

clinical

expertssuppliers

hospital

or other

clinical

center

Figure 11: Interoperability: systems get connected at all levels

Figure 11 shows that systems all levels in the hospital get connected. Connec-tivity is not sufficient to justify the integration cost, only if systems interoperate ina useful way true added value is created.

applicationscilinical analysis

clinical support

administrative

financial

workflow

vendorsPhilips

GE

Siemens

releasesR5

R6.2

R7.1

standardsDicom

HL7

XML

media, networksDVD+RW

memory stick

memory cards

bluetooth

11a/b/g

UTMS

languages

culturesUSA, UK,

China, India,

Japan, Korea

France, Germany

Italy, Mexico

based on multiple

delivered by multipleintegrating multiple in multiple

and multiple and multiple

Figure 12: Multi dimensional interoperability

An additional challenge for interoperability are the many dimensions in whichinteroperability is required, as shown in figure 12:



page: 7

• applications• languages, cultures• vendors• media, networks• standards• releases

4 Reliability

The amount of software (and technology) in products is increasing exponentially,as shown for televisions in Figure 13. Many other products show the same growthin the amount of software: Cardio Vascular X-ray systems, MRI scanners, wafer-steppers.

1965 1979

2000 1990

1 kB

64 kB2 MB

Moore's law

Fro

m: C

OP

A tu

toria

l, R

ob

va

n O

mm

erin

g

Figure 13: SW increase in televisions

Software is far from errorless. Studies of the density of errors in actual codeshow that 1000 lines of code typically contain 3 errors. In case of very goodsoftware processes and mature organizations this figure is 1 a 2 errors per kloc,in poor organizations it can be much worse.

Incremental increase of the code size will increase the number of hidden errorsin a system also exponentially, as shown in figure 141.

Note that the interoperability, as described in Section 3 increases the reliabilitythreat with a multiplying factor.

1Simple systems, which already contain fatal bugs are appearing: for instance software vendingmachines, which need a power down when the software is crashed.



page: 8

1000

10Mloc

100

1 Mloc

10

100 kloc

ma

nye

ars

an

d L

OC

(lin

es o

f co

de)

pe

r p

rod

uct

1990 1995 2000 2005

1000

10k

typ

ica

l a

mo

un

t o

f

err

ors

pe

r p

rod

uct

Based on average

3 errors/kloc

Figure 14: Increase of software threatens Reliability

5 Power consumption

Many technological improvements show an exponential increase: circuit density,storage capacity, processing power, network bandwidth. An exception is the energydensity in batteries, which is improved, but much less dramatically. The powerusage of processing per Watt has improved significantly, but unfortunately theprocessing needs have increased also rapidly. Overall the power consumption offor instance PC’s is more or less constant.

heat dissipation

stand by time

operational time

acoustic noise

power supply cost

faster

more

power

perform

ance

examples

data centers

insufficient power in

Amsterdam

MRI gradients: 66 mT/m

kWatts: cost, noise, heat

desktop

silent fanless

3G phone

standby time

operational time

wireless video

Figure 15: Power consumption and dissipation

Figure 15 shows these opposing forces: the need for less power consumptionand for more performance. There are multiple reasons to strive for less powerconsumption:

• less heat dissipation, easier transport of waste heat



page: 9

• increase stand by time• increase operational time• decrease acoustic noise of cooling• decrease power supply cost

Power consumption plays a role in all domains, for example in GSM phones(standby time, operational time, wireless video), data centers (Amsterdam cannotaccommodate more data centers, due to the availability of electrical power), MRIgradients (faster switching of higher gradient fields, amplifier costs, acoustics andcooling are significant technical challenge) and desktop PC’s (where a current trendis to silent fanless desktops).

6 Security

Several stakeholders have significant different security interests. Figure 16 shows3 categories with different interest and security solutions:

• Government and companies, which implement restrictive rules, which canbe rather privacy intrusive

• Consumers, who want to maintain privacy and at the same time usability ofservices2

• The content industry, who want to get fair payment for content creation anddistribution. Their solution is again very restrictive, even violating the rightof private copies, and characterized by a paranoia attitude: every customeris assumed to be a criminal pirate.

All stakeholders are confronted with threats: pirates, thieves, terrorists, dictators,et cetera. The challenge is to find solutions which respect all the needs, not onlythe needs of one of the stakeholders. Another challenge is to make systems suffi-ciently secure, where a little bit insecure quickly means entirely insecure. Last butnot least is the human factor often the weakest link in the security chain.

7 Creativity

The rapid technology development has changed the nature of product creationchallenges. Several decades ago the main challenge was to get the desired perfor-mance at all; Can we do it? This shifted to the problem of realizing the performancewithin practical constraints, such as cost, power and size; Can we make it?

The next challenge was to create the product with large amounts of peoplein a limited amount of time. Processes are a means to create products in large

2Which can be quite conflicting. Who may have access to your medical data? Do you still wantto control this access in case of emergencies?



page: 10

company

government

security

restrictive

intrusive

content

industry

digital rights

restrictive

paranoia

co

nsu

me

rs

pri

va

cy

us

ab

ilit

y

freedom

protection

conflicting interests

pira

tes

thie

ve

s

terr

orists

dic

tato

rs

threats

Figure 16: Security conflicting interests

Performance

Cost,

Power, Size

People,

Process

Imagination

Can we

do it?

Can we

make it?

Can we

organize it?

Can we

conceive it?

Time

Am

bitio

n

leve

l

from: Ad Huijser Philips Software Conference 2001

Figure 17: Creativity as limiting factor

teams; Can we organize it? Today many well defined applications and servicescan be realized, the real problem has shifted to which applications and functionsare needed? Which services can be used to sustain a business? Can we conceiveit?

8 The role of the Embedded Systems Institute

The objective of embedded systems institute is to build up, consolidate and transferknowledge how to create embedded systems effectively. Figure 18 summarizes therole of the institute by means of an annotated sentence.

To create embedded systems means and methods are needed to specify, design,



page: 11

power, cost, economy

skills, legislation

performance

interoperability

productivity

reliability

methods to:

specify, design,

test and verify;

f.i. modeling

How to create embedded systems

which satisfy the functionality, and quality needs

and which fit in the limiting constraints

Figure 18: Role of Embedded Systems Institute ESI

test and verify, for instance modelling methods. These embedded systems mustfulfil the functionality and the quality needs.

Embedded systems are mostly created in an industrial setting, which means thatmany practical constraints are present during the product creation. The creationmethods must cope in a practical way with these constraints. The EmbeddedSystems Institute cooperates closely with an industrial partner, the so-called industryas laboratory approach.

References

[1] Gerrit Muller. The system architecture homepage. http://www.gaudisite.nl/index.html, 1999.

[2] Gerrit Muller. Light weight architectures; the way of the future? http://www.gaudisite.nl/info/LightWeightArchitecting.info.html, 2001.

[3] Gerrit Muller. Challenges in high-tech, illustrated in the medical domain.http://www.gaudisite.nl/DYOFPaper.pdf, 2003.

HistoryVersion: 1.0, date: May 19, 2004 changed by: Gerrit Muller



page: 12

http://www.gaudisite.nl/index.html

http://www.gaudisite.nl/index.html

http://www.gaudisite.nl/info/LightWeightArchitecting.info.html



http://www.gaudisite.nl/DYOFPaper.pdf

• Added article versionVersion: 0, date: April 5, 2003 changed by: Gerrit Muller

• Created, no changelog yet



page: 13

Documents

Opportunities and challenges in embedded systems