1 Vad hnder i dag? Kur splan Kur sdelt agar na ska f r vr va - - PDF document

▶
1
SMART_READER_LITE
LIVE PREVIEW

1 Vad hnder i dag? Kur splan Kur sdelt agar na ska f r vr va - - PDF document

Feb 04, 2023 634 likes •738 views

Def ining Dist r ibut ed Syst ems A dist ribut ed syst em is one in which component s locat ed at net worked Dist r ibuer ade syst em comput ers communicat e and coordinat e t heir act ions only by passing messages C-kurs Om

slide-1
SLIDE 1

1

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 1

Dist r ibuer ade syst em

C-kurs 5 poäng HT-02

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 2

Def ining Dist r ibut ed Syst ems

  • ”A dist ribut ed syst em is one in which

component s locat ed at net worked comput ers communicat e and coordinat e t heir act ions only by passing messages”

  • ”Om en maskin som du aldrig hört t alas om

går ner och det hindrar dig f r ån at t j obba så har du et t dist ribuerat syst em”

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 3

Exemples of DS: I nt er net

intranet ISP desktop computer: backbone satellite link server: ☎ network link: ☎ ☎ ☎ 3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 4 the rest of email server Web server

Desktop computers

File server router/firewall print and other servers

  • ther servers

print Local area network email server the Internet

Exemples of DS: I nt r anet

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 5

Exemples of DS: Mobile Comput ing

Laptop Mobile Printer Camera Internet Host intranet Home intranet WAP Wireless LAN phone gateway Host site 3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 6

Consequences of DS

  • Concurr ency
  • No global clock
  • I ndependent f ailures
slide-2
SLIDE 2

2

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 7

Vad händer i dag?

  • Kursledning
  • Kursplan
  • Förra år et
  • Året s kurs
  • Regist rering
  • I nt rodukt ion t ill Dist ribuerade syst em

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 8

Kur splan

  • Kur sdelt agar na ska f ör vär va

– kunskaper om t eoret iska modeller f ör dist ribuer ade syst em – kunskaper om problem och lösningar i samband med design och konst rukt ion av dist ribuerade syst em

  • Kur sen behandlar:

– Arkit ekt urmodeller f ör dist r ibuerade syst em – Klient -Server, peer-t o-peer, t r ansakt ioner, t ransparens, namngivning, f elhant ering, resursf ördelning och synkroniser ing – Dat asäkerhet saspekt er i et t bret t per spekt iv – Met oder och t ekniker f ör dat asäker het – Dist r ibuerad progr ammer ing och ”middleware”

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 9

Kur splan (f or t s)

  • Moment 1, t eoridel, 3 poäng

– Teor ier , met oder , algor it mer och pr inciper

  • Moment 2, laborat ionsdel, 2 poäng

– Obligat or iska uppgif t er

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 10

För r a år et

  • Flera f öreläsare
  • Kursdisposit ion – Bra, men spridda åsikt er
  • Gruppövningar – Den var bra
  • Kurslit t erat ur – Bra, men f ormell och

t ungläst

  • Koppling t eori och labbar – Okej , blandat
  • Tyngd på labbar – Lagom
  • Lit e högre arbet sbelast ning än snit t et

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 11

År et s kur s

  • Kurslit t erat ur
  • Fokus på

– Middlewar e – progr ammer ing – Säker het – Dist r ibuerade algor it mer

  • Upplägg

– Föreläsningar – Gr uppövningar – Labor at ioner

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 12

Labor at ioner

  • 3 st ycken

– Debat t om säker het – E-handelssyst em, huvudlabben – En lit en om EJ B

  • Tid t ill f örf ogande
  • Grupper om 2 eller 3, gäller ej debat t en
  • Koppling t ill t eorin
  • MA336
slide-3
SLIDE 3

3

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 13

Under visning

  • P

edagogisk t anke

  • Val av present at ionshj älpmedel
  • OH-bilder

– Källa – I nnehåll – Spr åk

  • Vad bör man lära sig?

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 14

The r est of t oday

  • Challenges
  • Archit ect ural models
  • Fundament al models

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 15

Resour ce shar ing: The Web

Internet Browsers Web servers www.google.com www.cdk3.net www.w3c.org Protocols Activity.html http://www.w3c.org/Protocols/Activity.html http://www.google.comlsearch?q=kindberg http://www.cdk3.net/ File system of www.w3c.org

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 16

Dist r ibut ed Syst ems

  • What is dist r ibut ed?

– Dat a – Comput ing power – Services – Ent it ies – Securit y – Name space – Time

  • Challenges

– Het erogeneit y – Openness – Securit y – Scalabilit y – Failur e handling – Concurr ency – Transparency

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 17

Het er ogeneit y

  • Applies t o

– Net works – Comput er hardwar e – Operat ing syst ems – Progr amming languages – I mplement at ions by dif f erent developers

  • Middlewar e - sof t war e

layer t hat provides pr ogr amming abst r act ion and mask het erogeneit y

  • Mobile code - code

t hat can be sent f r om

  • ne comput er t o

anot her

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 18

Openness

  • Ext ended and re-

implement ed

  • Ver y impor t ant in t he

indust r y

  • RFC (r equest s f or

comment s)

– You’ll use it in your

  • work. I n t his course.

And probably in t he rest of your lif e

  • Open syst ems;

– Key int erf aces are published

  • Open dist r ibut ed

syst ems can be

– const ruct ed f rom het erogeneous hardware and sof t ware.

slide-4
SLIDE 4

4

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 19

Scalabilit y

  • Cont r olling t he cost of physical r esour ses
  • Cont r olling t he perf or mance loss
  • P

r event ing sof t war e resour se r unning out

  • Avoiding per f ormance bot t lenecks

Date Computers Web servers 1979, Dec. 188 1989, July 130,000 1999, July 56,218,000 5,560,866

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 20

Scalabilit y

Date Computers Web servers Percentage 1993, July 1,776,000 130 0.008 1995, July 6,642,000 23,500 0.4 1997, July 19,540,000 1,203,096 6 1999, July 56,218,000 6,598,697 12

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 21

Failur e handling

  • Techniques f or dealing wit h f ailures

– Det ect ing f ailures – Masking f ailur es – Tolerat ing f ailur es – Recovery f rom f ailures – Redundancy

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 22

Tr anspar ency

Access transparency: enables local and remote resources to be accessed using identical

  • perations.

Location transparency: enables resources to be accessed without knowledge of their location. Concurrency transparency: enables several processes to operate concurrently using shared resources without interference between them. Replication transparency: enables multiple instances of resources to be used to increase reliability and performance without knowledge of the replicas by users or application programmers. Failure transparency: enables the concealment of faults, allowing users and application programs to complete their tasks despite the failure of hardware or software components. Mobility transparency: allows the movement of resources and clients within a system without affecting the operation of users or programs. Performance transparency: allows the system to be reconfigured to improve performance as loads vary. Scaling transparency: allows the system and applications to expand in scale without change to the system structure or the application algorithms.

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 23

Syst em Models

  • Ar chit ect ur al models

– Concerned wit h t he placement s of it s par t s and t he relat ionships bet ween t hem. – Examples

  • Client -ser ver
  • Peer -t o-peer
  • Fundament al models

– Concerned wit h a more f ormal descript ion of t he propert ies t hat are common in ar chit et ur al models – Models

  • I nt er act ion model
  • Failur e model
  • Secur it y model

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 24

Sof t war e and har dwar e layer s

Applications, services Computer and network hardware Platform Operating system Middleware

slide-5
SLIDE 5

5

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 25

Client -ser ver

Server Client Client invocation result Server invocation result Process: Key: Computer:

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 26

Mult iple - ser ver s

Server Server Server Service Client Client

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 27

Web pr oxy ser ver

Client Proxy Web server Web server server Client

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 28

Peer pr ocesses

Coordination Application code Coordination Application code Coordination Application code 3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 29

Mobile code - web applet s

a) client request results in the downloading of applet code Web server Client Web server Applet Applet code Client b) client interacts with the applet

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 30

Thin client s

Thin Client Application Process Network computer or PC Compute server network

slide-6
SLIDE 6

6

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 31

Mobile devices and spont aneous net wor king

Internet gateway PDA service Music service service Discovery Alarm Camera Guests devices Laptop TV/PC Hotel wireless network

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 32

Mobile devices and spont aneous net wor king (cont .)

  • Key f eat ures and ot her issues

– Easy connect ion t o a local net wor k – Easy int egr at ion wit h local ser vices – Ot her issues

  • Limit ed connect ivit y
  • Securit y and privacy

– Discover y ser vices

  • Regist r at ion
  • Lookup services

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 33

Desgn r equir ement s f or dist r ibut ed ar chit ect ur es

  • P

erf ormance issues

  • Qualit y-of -service
  • Use-of -caching and replicat ion
  • Depedandabilit y issues

– Fault t oler ance – Secur it y

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 34

Fundament al models

  • Quest ions

– Main ent it ies? – How do t hey int eract ? – Behaviour?

  • P

ur pose

– Make all relevant assumt ions explicit – Make gener alizat ion

  • I nt er act ion
  • Failur e
  • Secur it y

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 35

I nt er act ion model

  • Two signif icant f act ors af f ect ing

int eract ion

– Communicat ion per f or mance is of t en a limit ing char act er ist ic. – I t is impossimble t o maint ain a single not ion of t ime.

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 36

Per f or mance of communicat ion channels

  • Lat ency
  • Bandwidt h
  • J it t er
slide-7
SLIDE 7

7

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 37

Event or der ing

send receive send receive m1 m2 2 1 3 4 X Y Z Physical time A m3 receive receive send receive receive receive t1 t2 t3 receive receive m2 m1

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 38 process

p

process q Communication channel

send

Outgoing message buffer Incoming message buffer

receive m

Communicat ion omission f ailur es

  • Omission and ar bit rar y f ailur es

– Fail-st op – Crash – Omission – Send-omission – Receive-omission – Ar bit rary (Byzant ine)

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 39

Timing f ailur es

Class of Failure Affects Description Clock Process Process’s local clock exceeds the bounds on its rate of drift from real time. Performance Process Process exceeds the bounds on the interval between two steps. Performance Channel A message’s transmission takes longer than the stated bound.

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 40

Reliabilit y of one-t o-one communicat ion

  • Validit y
  • I nt egrit y
  • Threat s

– Ret r ansmit t ing – Malicious users

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 41

Secur it y models: Pr ot ect ing

  • bj ect s

Network invocation result Client Server Principal (user) Principal (server) Object Access rights

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 42

The enenmy

  • Threat s t o processes
  • Threat s t o communicat ion channels

Communication channel Copy of

m

Process

p

Process

q m

The enemy

m’

slide-8
SLIDE 8

8

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 43

Def eat ing secur it y t hr eat s

  • Cr ypt ography and shared secr et s
  • Aut hent icat ion
  • Secur e channels

Principal

A

Secure channel Process

p

Process

q

Principal

B

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 44

Summar y

  • I nt roduct ion t o t he course
  • What is a dist ribut ed syst em

– ”A dist r ibut ed syst em is one in which component s locat ed at net worked comput ers communicat e and coor dinat e t heir act ions only by passing messages”

  • Consequences

– Concur r ency – No global clock – I ndependent f ailures

3/9 - 02 Distribuerade system - Jonny Pettersson, UmU 45

Summar y (cont .)

  • Challenges

– Het erogeneit y – Openness – Securit y – Scalabilit y – Failur e handling – Concurr ency – Transparency

  • Ar chit ect ur al models
  • Fundament al models