Transcript a client

Revision
Distributed Systems Introduction
Distributed Systems
 A distributed system is:
Distributed system is one in which components located at
networked computers communicated and coordinate their
actions only by passing message –G. Coulouris
 A collection of independent computers that appears to its
users as a single coherent system. - S. Tanenbaum
 It leads to concurrency of components, lack of a global clock
and independent failures of components
 - Characteristics of distributed systems
 Concurrency
 No Global clock
 Independent failures
Why Distributed Systems?
 Main features
Geographical distribution of autonomous computers
Communication through cable/fiber/wireless/...connections
A collection of independent computers that appears to its
users as a single coherent system logically( called a single
view system).
Advantages
interaction, co-operation, and sharing of resources
Benefits
reduced costs, improved availability and performance
Scalability, resource sharing, fault tolerance.
3
Distributed Systems
Data
0.
-
Distributed system ?
Autonomous systems
Computer network
Distributed software
Communication
Network
Data
0.
-
Important consequences
Concurrency
No global clock
Independent failure
4
Distributed Systems- consequences
 Concurrency
 web pages /
sharing resources
 No Global clock
No single global notion of the correct time
 Independent failures
 slow network /unexpected termination
 each component of the system can fail independently
Introduction of Distributed Systems
 Definitions of Distributed Systems
A collection of independent computers that appears to its
users as a single coherent system.
A distributed system organized as middleware.
 Note that the middleware layer extends over multiple machines.
6
Importance of Distributed Computing
 Distributed (computer) systems are critical for functioning of
many organizations
Banks
Transport
Telecommunications
 Distributed Application
A set of processes that are distributed across a network of
machines and work together as an ensemble to solve a common
problem
7
Typical examples
 The Internet
 global network of interconnected computers which communicate through IP
protocols
 A vast interconnected collection of computer networks of many different types.
 An Intranets
 a separately administered network with a boundary that allows to enforce
local security policies
 A portion of the Internet. Router/firewall exclusive File services/ Impeding
Firewalls/The cost of installation
 Mobile and ubiquitous computing
 laptops, PDAs, mobile phones, printers, home devices, ...
 technological advance in device miniaturization and wireless networking.
 World-Wide Web
 system for publishing and accessing resources and services across the Internet
 HTML/URLs/HTTP/Dynamic Pages
8
Distributed Systems’ Challenges
 Challenges
 Heterogeneity
 Openness
 Security
 Scalability
 Failure handling
 Concurrency
 Transparency
 Due to:
 complexity
 size
 changing technologies
 society’s dependence
Distributed Systems’ Challenges
 Heterogeneity
 Networks, hardware, OSs, P-languages…etc.
 Solution-Protocol, middleware.
 Openness
 Cumbersome & slow moving - follow standardization
 Security
 Hospital/E-commerce/banking
 Failure handling
 Software/hardware - proper policies
 Concurrency
 Sharing resource at the same time - operation should be synchronized
 Scalability / Transparency
Heterogeneity
 varying software and hardware
OSs, networks, computer hardware, program languages,
implementations by different developers
need for standards of protocols, middleware
 Heterogeneity and mobile code support
virtual machine approach (cf, Java applets)
11
Openness
 independence of vendors
 publishable key interfaces
CORBA(Common Object Request Broker Architecture)
 publishable communication mechanisms
Java RMI(Remote Method Invocation)
12
Security
 confidentiality (protect against disclosure)
 cf, medical records
 integrity (protect against alteration and interference)
 cf, financial data
 Need encryption and knowledge of identity
1. Denial of Service attacks->Distributed DoS
2. Security of mobile code
13
Scalability
 Design of scalable distributed systems
Controlling the cost of physical resource
Controlling the performance loss Preventing software resource
running out
Avoiding performance bottleneck
Examples of scalability limitations
Concept
Example
Centralized services
A single server for all users
Centralized data
A single on-line telephone book
Centralized algorithms
Doing routing based on complete
information
14
Scalability Techniques (1)
 The difference between letting: No. of servers or clients
a) a server or
b) a client check forms as they are being filled
Thin
Thick(=fat)
Thick(=fat)
Thin
15
Scalability
Characteristics of decentralized algorithms:
No machine has complete information about the
system state.
Machines make decisions based only on local
information.
Failure of one machine does not ruin the algorithm.
There is no implicit assumption that a global clock
exists.
Failure handling
Ability to continue computation in the presence of failures.
 Detecting failures
 Masking failures(= hiding failure)
 Tolerate failures
 Recovery from failures
 Redundancy
17
Concurrency
Processes execute simultaneously and share resources.
 synchronization
 inter-process communication(IPC)
18
Transparency
Concealment of the separated nature of system from
user/programmer
=>Network transparency
Access transparency + Location Transparency
cf .log on, email, …. on network
 Transparencies=> By ANSA Reference Manual & ISO Reference
Model for Open Distributed Processing (RM-ODP)
19
Transparencies(1) G. Coulouris
•Access transparency: enables local and remote resources to be
accessed using identical operations.
•Location transparency: enables resources to be accessed without
knowledge of their physical or network location (for example, which
building or IP address).
•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.
Transparencies(2) G. Coulouris
•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.
Distributed Systems : System Models
22
Application Layering (1)
Recall previously mentioned layers of
architectural style
The user-interface level
The processing level
The data level
Application Layering (2)
 The simplified organization of an Internet search engine
into three different layers.
Multitiered Architectures (1)
The simplest organization is to have
only two types of machines:
A client machine containing only the
programs implementing (part of) the userinterface level
A server machine containing the rest,
the programs implementing the processing and data
level
Multitiered Architectures (2)
 Alternative client-server organizations (a)–(e).
Client-Sever System ctd
 Client-Server Application
DS.
 Client : Process that request a service
 Server : Process that provides a service
 Client blocks until server responds
 Client is invoked by end users when they require a service
 Server waits for incoming requests
 Server can have many clients making concurrent requests
 Server is a program with special system privileges
27
Client-Sever System – Applying Database system ctd
“Gartner Group” - 2-tier Database system’s architecture
user interface = presentation, application = logic, database =
data
Client
system
Server
system
X-Window
Telnet
World Wide Web(html, java–CGI http server)
Network File(DB) systems
Distributed databases
28
Client-Sever System ctd
 Gartner Group’ Configuration(2-tier, 3-tier Architectures)
Client
system
Server
system





Distributed data : Distributed databases
Remote data : Network File systems
Distributed transaction : World Wide Web(html, java – CGI http server)
Remote presentation : Telnet
Distributed presentation : X-Window
29
Software Layers
language and run-time
support for program
Interactions
conventional
and
distributed
applications
applications
extended services
available to those of
the distributed system
Open (distributed) services
Middleware( CORBA, Java RMI, Web service, DCOM, RD-ODP)
responsible for basic local
Operating system
resource management
(memory allocation/protection,
process creation and
scheduling, local inter-process
communication and peripheral
devices handling)
Computer and network (TCP.UDP/IP). hardware
Platform(green part)
30
Software layers ctd
 Service layers
 Higher-level access services at lower layers
 Services can be located on different computers
 Process types:
server process
client process
peer processes(client process + server process)
 Ex. remote conferences
31
Important layers
 Platform
lowest-level hardware+software( OS + Communication)
common programming interface
different implementations of operating system facilities
for co-ordination & communication
 Middleware
programming support for distributed computing
32
Software Concepts
 DOS (Distributed Operating Systems)
 NOS (Network Operating Systems)
 Middleware
System
Description
Main Goal
DOS
Tightly-coupled operating system for multiprocessors and homogeneous multicomputers
Hide and manage
hardware resources
NOS
Loosely-coupled operating system for heterogeneous Offer local services to
multicomputers (LAN and WAN)
remote clients
Middleware
Additional layer atop of NOS implementing generalpurpose services
Provide distribution
transparency
Network Operating System
 OSes can be different (Windows or Linux)
 Typical services: rlogin, rcp
 Fairly primitive way to share files
Distributed Operating Systems
 But no longer have shared memory
 Provide message passing
 Can try to provide distributed shared memory
 But tough to get acceptable performance
Distributed System as Middleware
Middleware
 Definitions and Examples of Middleware
37
Middleware provides...
ctd
 Support for distributed processes/objects:
suitable for applications programming
communication via the following mechanisms;
 remote method invocation (Java RMI), or
 remote procedure call (Sun RPC)
 Services infrastructure for application programs
naming, security, transactions, event notification, ...
Commercial products: CORBA, DCOM,…
38
Middleware and Openness
1.23
 In an open middleware-based distributed system, the protocols
used by each middleware layer should be the same, as well as
the interfaces they offer to applications.
 If different, there will be compatibility issues
 If incomplete, then users will build their own or use lowerlayer services (frowned upon)
Main types of Architecture models
* Client-server architecture and Peer-to-peer architecture
 Client-server model
first and most commonly used
 Multiple server model
to improve performance and reliability
e.g. search engines (more than 1000’s of computers)
 Proxy server model
to reduce load on network,
to provide access through firewall
 Peer process model
when faster interactive response needed
40
Client-Server Model
Client
invocation
Server
result
Backed server
result
invocation
Server
Front server
Client
Key:
Process:
Computer:
Front server acts as a client for backed servers
41
Service Provided by Multiple server Model
Service
Server
Client
Server
Replicated object, called
Replica
Client
Server
- Performance, reliability
Servers -may
interact
Replicated
and distributed database
42
Proxy Server Model and Caching Mechanism
Web
server
Client
Proxy
server
Web
server
Client
Caching mechanism
43
Peer Process Model
Application
Application
Coordination
code
Coordination
code
Application
Coordination
code
when faster interactive response needed
44
Horizontal Distribution
 Distribute servers across nodes
E.g., Web server for load balancing
 Distribute clients in peer-to-peer systems.
Client-server and mobility
 Mobile code
downloaded from server, runs on locally
e.g. web applets
 Mobile agent (code + data)
travels from computer to another
collects information, returning to origin
=> Beware! Security risks
46
Web applets
 Client requests results, applet code is downloaded;
Client
Applet code
Web
server
 Client interacts with applet;
Client
Applet
Web
server
47
Thin & Fat Clients
 Thin client = Network computer
Typically no local storage
 has only presentation function(= user interface)
 Fat Client = Desktop PC, Workstation
has both presentation function and logic function
(=application)
 Motivation for Thin Clients
Hidden costs of System administration and supports
 Networks computers move toward centralized system Admin. , but local
processing at client
Java(mobile code)  an enabling technology
 E.g. PDA
48
Design Requirements for Distributed Architectures
Judging how good the architecture is...
 1)Performance
how fast will it respond?
 2)Quality of Service(QoS)
are video frames and sound synchronized?
 3)Dependability
does it work correctly?
49
1) Performance
 Responsiveness
fast interactive response delayed by remote requests
use of caching, replication
 Throughput
dependent on speed of server and data transfer
 Load balancing
use of applets, multiple servers
50
Caching and Replication
 Caches and Web Proxy Servers(Proxies)
 Cache consistency protocol in Chapter 8. Distributed File System
 Web-caching protocol in HTTP protocol
Web
server
Client
Proxy
server
Replicated object
Client
caching
Web
server
called,
replica
Browsers and proxies can validate a cached response
 If fails, web server returns a fresh response instead of stale response
Browsers and proxies stores expiry time of cached resource and
server time with cached responses
51
2) Quality of Service (QoS)
Non-functional properties experienced by users:
 Deadline properties
hard deadlines (must be met within T time units)
soft deadlines (`there is a 90% chance that the video frame will
be delivered within T time units, ex, 10 units -90% 9units)
 multimedia traffic, video/sound synchronization
 depend on availability of sufficient resources
 Adaptability
ability to adapt to changing system configuration
52
3) Dependability
 Correctness
Ensuring correctness of distributed and concurrent
programs
e.g. use of verification
 Fault-tolerance
ability to tolerate/recover from faults
e.g. use of redundancy
 Security
ability to withstand malicious attack
e.g. use of encryption, etc
53
Fundamental Models for Distributed Systems
 Sharing some fundamental properties
processes that communicate with one another by sending msgs
 Essential ingredients of System model
what are the main entities in the systems ?
how do they interact ?
what are characteristics that affect their individual and collective
behavior ?
 Purpose of system model
to make explicit relevant assumptions at system modeling
to generate the concerning possible or impossible assumptions
54
3 Fundamental Models ctd
 Discussions of fundamental models
Interaction Model
Failure Model
Security Model (skipped in this review)
55
Interaction Model
 Distributed systems
have many processes, complex interactions among them
 transmission of message between processes(local, remote)
 Interacting processing
communication performance
impossible to maintain global time( each system has its local
time)
 Performance of communication channel
Latency(message, network, system)
bandwidth
Jitter
56
Interaction Model ctd
 Computer clocks and timing events
clock drift rate(Global time needed)
clock synchronization
 Two variants of the interaction model
Synchronous DS has on;
 process is executing in a known lower/ upper bounded time
 message is received within a known bounded time
 known local clock’s drift rate
Asynchronous DS has no bounds on;
 process execution speed
 message transmission delay
 clock drift rate
-so, not suitable for multimedia service
57
Failure Model
 In DS, processes and channels may failure
3 failures : Omission, Arbitrary, timing failures
 Omission failures
process omission failures
 Correct, fail-stop, timeouts
communication failures
 send/receive omission failures
process
p
process
send
m
q
Receive
Communication channel
Outgoing message buffer
Incoming message buffer
58
Failure Model ctd
Omission and arbitrary failures
Class of failure
Fail-stop
Affects
Process
Crash
Process
Omission
Channel
Send-omission
Process
Receive-omission
Process
Arbitrary
(Byzantine)
Process or
channel
Description
Process halts and remains halted. Other processes may
detect this state.
Process halts and remains halted. Other processes may
not be able to detect this state.
A message inserted in an outgoing message buffer never
arrives at the other end’s incoming message buffer.
A process completes as
end, but the message is not put
in its outgoing message buffer.
A message is put in a process’s incoming message
buffer, but that process does not receive it.
Process/channel exhibits arbitrary behaviour: it may
send/transmit arbitrary messages( with wrong values)
at arbitrary times, commit omissions;
a process may stop or take an incorrect step.
59
Failure Model ctd
 Arbitrary failures(=Byzantine failure, Data failure)
 Process/channel exhibits arbitrary behaviour(delivering corrupted, nonexisted message )
 Timing Failures
Class of Failure
Clock
Affects
Process
Description
Performance
Process
Process exceeds the bounds on the interval
Performance
Channel
Process’s local clock exceeds the bounds on its
rate of drift from real time.
between two steps.
A message’s transmission takes longer than the
stated bound.
 The others;
 Masking failures
 Checksum failure
 Failure of reliability of one-to-one communication
 Threats of validity and integrity
60
Networking and Interworking in DS
Inter-Process Communication
Operating System Support
Dr. Sunny Jeong. [email protected]
With Thanks to Prof. G. Coulouris,
Prof. A.S. Tanenbaum and Prof. S.C Joo
61
Network issues in DSs
 Performance
Latency
Data transfer rate
 Message transfer rate time = latency + a massage length/ data transfer rate
 Total system bandwidth of network
 Throughput in the end systems
 total volume of traffic can be transferred across network in a time
 Scalability
No designable to cope with size and load about network growing
 Reliability
recoverable from communication failures
62
Network issues in DSs ctd
 Security
 protecting network and computers, ex) firewall between intranet and
internet
 Mobility
portability of computer and handled digital devices using
wireless network
location and identification are depicted with each other
no designable to cope with size and load about network growing
 QoS(Quality of Service)
guarantee for requirements of computer and network
 to meet deadline, bandwidth, bounded latency
 Multicasting
 One-to-many communication
63
Network principles
 Mode of transmission
 Switching schemes
 Protocol suites
 Routing
 Congestion control
64
Mode of transmission
 Packets
messages divided into packets( on Transport Layer)
packets queued in buffers before sent onto link
QoS not guaranteed
 Data streaming
links guarantee QoS (rate of delivery)
for multimedia traffic
need higher bandwidth
65
Switching schemes
 Broadcasts (Ethernet, wireless)
send messages to all nodes
nodes listen for own messages (carrier sensing)
 Circuit switching (phone networks)
 Packet switching (TCP/IP)
store-and-forward
unpredictable delays
 Frame/cell relay (ATM)
bandwidth & latency guaranteed (virtual path)
small, fixed size packets (padded if necessary)
 53bytes= header 5 + body 48
avoids error checking at nodes (use reliable links)
66
Protocols ( ISO Open System Interconnection view)
(Trivial File Transfer
Protocol)
Internet Control(Group) Message Protocol
(Reverse Address Resolution Protocol)
67
OSI protocol summary
Layer
Application
Description
Protocols that are designed to meet the communication requirements of
specific applications, often defining the interface to a service.
Protocols at this level transmit data in a network representation that is
independent of the representations used in individual computers, which may
differ. Encryption is also performed in this layer, if required.
At this level reliability and adaptation are performed, such as detection of
failures and automatic recovery.
Examples
HTTP, FTP , SMTP,
CORBA IIOP
Secure Sockets
(SSL),CORBA Data
Rep.
Transport
This is the lowest level at which messages (rather than packets) are handled.
Messages are addressed to communication ports attached to processes,
Protocols in this layer may be connection-oriented, or connectionless.
TCP, UDP
Network
Transfers data packets between computers in a specific network. In a WAN
or an internetwork this involves the generation of a route passing through
routers. In a single LAN, no routing is required.
Responsible for transmission of packets between nodes that are directly
connected by a physical link. In a WAN transmission is between pairs of
routers or between routers and hosts. In a LAN it is between any pair of hosts.
The circuits and hardware that drive the network. It transmits sequences of
binary data by analogue signalling, using amplitude or frequency modulation
of electrical signals (on cable circuits), light signals (on fibre optic circuits)
or other electromagnetic signals (on radio and microwave circuits).
IP, ATM virtual
circuits
Presentation
Session
Data link
Physical
Ethernet MAC,
ATM cell transfer,
PPP
Ethernet base- band
signalling, ISDN
68
Internetwork protocol
 Intenetwork layer(=Virtual network layer)
 internet packet  destination (by datagram protocol)
 Network interfaces layer
Internetwork packets  suitable packets  underlying layer
Message
 Underlying network layer Layers
Application
Internetwork
protocols
Transport
Internetwork
Internetwork packets
Network interface
Network-specific packets
Underlying
network
protocols
Underlying network
69
Port & Addressing
 Port
network-independent message transport service
between networks ports
software-definable destination points for
communications
 Addressing
delivering messages to destination with transport
addresses
Transport address
 Network address + port number
70
Packet delivery
 In network layer
datagram packet delivery(IP in Ethernet, most wired
and wireless LAN technologies)
virtual circuit packet delivery(ATM)
 In transport layer
connection-oriented transmission(TCP)
 Reliable communication with static routing table(ISO, X.25)
 Ex) remote login(Telnet), FTP, HTTP(big-sized file), stream data
connectionless transmission(UDP)
 Unreliable communication with pre-defined routing table
 Ex) rcp, rwho, RPC, HTTP(small-sized file), FTP(non-error bulk file)
71
Routing
 Necessary in non-broadcast networks (cf Internet) : Hop
by Hop
 Distance-vector algorithm for each node
stores table of state & cost information of links, cost infinity for
faulty links
determines route taken by packet (the next hop)
periodically updates the table and sends to neighbors
may converge slowly [Bellman-Ford]
 RIP-1(Router Information Protocol) for Internet
Local router table changes
use default routes, plus multicast and authentication
better convergence( routes better route to an existing destination)
72
Congestion control
 When load on network exceeds 80% of its capacity
packet queues long, links blocked
 Solutions(in datagram-based network layers)
packet dropping
 reliable of delivery at higher levels
reduce rate of transmission
 nodes send choke packets (Ethernet)
 special message requesting a reduction in transmission rate
 transmission control (TCP)
transmit congestion information to each node
 QoS guarantees (ATM)
73
Transport level protocols
 UDP (basic, used for some IP functions)




uses IP address+port number
no guarantee of delivery, optional checksum
messages up to 64KB
Connectionless transmission(  Unreliable and Asynchrnous communication with
pre-defined routing table)
 Datagram service
 Ex) rcp, rwho, RPC, HTTP(small sized), FTP(non-error bulk file)
 TCP (more sophisticated, most IP functions)






data stream abstraction, reliable delivery of all data
messages divided into segments, sequence numbers
sliding window, acknowledgement+retransmission
buffering (with timeout for interactive applications)
checksum (if no match segment dropped)
Connection-oriented transmission(  Reliable and Synchronous communication with
static routing table(ISO, X.25))
 Stream service
 Ex) remote login(Telnet), FTP, HTTP(bulk file), stream data
74
IP(TCP/IP) Addressing
 IP Address(IPv4)
75
IP(TCP/IP) Addressing ctd
 IP Structure(Universal)(IPv4 : 4bytes: 32bits)
7
Class A:
0
24
Network ID
Host ID
14
Class B:
1
0
16
Network ID
Host ID
21
Class C:
1
1 0
8
Network ID
Host ID
28
Class D (multicast):
1
1 1
0
Multicast address
27
Class E (reserved):
1
1 1
1 0
unused
IP Packet layout
header
IP addres s of s ource
IP addres s of des tination
data
up to 64 kiloby tes
76
IP(TCP/IP) Addressing ctd
octet 1
octet 2
Network ID
Class A:
1 to 127
Class C:
Range of addresses
Host ID
0 to 255
0 to 255
1.0.0.0 to
127.255.255.255
0 to 255
0 to 255
128.0.0.0 to
191.255.255.255
0 to 255
Host ID
1 to 254
0 to 255
Network ID
Class B:
octet 4
octet 3
Host ID
128 to 191
0 to 255
192 to 223
Network ID
0 to 255
192.0.0.0 to
223.255.255.255
Multicast address
Class D (multicast):
224 to 239
0 to 255
0 to 255
1 to 254
224.0.0.0 to
239.255.255.255
Class E (reserved):
240 to 255
0 to 255
0 to 255
1 to 254
128.0.0.0 to
247.255.255.255
•
•
•
•
Address
Address
Address
Address
194.0.0.0
198.0.0.0
200.0.0.0
202.0.0.0
to
to
to
to
195.255.255.255.are
199.255.255.255.are
201.255.255.255.are
195.203.255.255.are
in
in
in
in
Europe
N. America
Central & South America
Asia and Pacific
77
IPv4-> IPv6
 IPv6 Address
• Large Address space - 128 bit addresses
–Every toaster can have its own IP address
• Aggregation-based address hierarchy
–Efficient backbone routing
• Efficient and Extensible IP datagram
–No fragmentation by routers
–64 bits field alignment
–Simpler basic header
• Auto-configuration
• Security
• IP Renumbering part of the protocol
78
IPv4-> IPv6
 3FFE:0B00:0C18:0001:0290:27FF:FE17:FC0F
TLA
NLA(s)
16 bits
32 bits
SLA
16 bits
Interface ID
64 bits
 TLA – top level aggregator
 Primary providers
 NLA: Next Level Aggregator
 Can have multiple NLA as sub-NLA
 SLA: Site Level Aggregator
 Your site (16 bits)
 Addresses are allocated from your provider
 If you change provider, your prefix changes
 But renumbering (of hosts, routers and sites) has been included in the IPv6 protocol
79
IPv4-> IPv6
 IPv6 Header layout(16bytes : 128bits)
Version (4 bits)
Priority (4 bits)
Flow label (24 bits)
Payload length (16 bits)
Next header (8 bits)
Hop limit (8 bits)
Source address
(128 bits)
Destination address
(128 bits)
IPv6’s main advances(Adapted by IETF in 1994)
address space(2128 = 3×1038 IPs), routing speed up
Real-time and other special services
Future evolution
Multicating & anycasting
security
80
MobileIP
 At home normal, when elsewhere mobile host:
notifies HA(Home Agent) before leaving
informs FA(Foreign Agent), who allocates temporary care-of IP
address & tells HA
 Packets for mobile host(MH):
first packet routed to HA, encapsulated in MobileIP packet and
sent to FA (tunnelling)
FA unpacks MobileIP packet and sends to mobile host
sender notified of the care-of address for future communications
which can be direct via FA
 Problems
efficiency low, need to notify HA
81
MobileIP routing
Sender
4.Subsequent IP packets
tunnelled to FA
Mobile host (MH)
3.Address of FA
returned to sender
1.First IP packet
addressed to MH
Internet
Foreign agent(FA)
Home
Agent
HA
2.First IP packet
tunnelled to FA
82
Asynchronous Transfer Mode(ATM)
 Asynchronous Transfer Mode(ATM)
Multimedia data(voice and video), distributed system services
Packet switching network based on Cell-relay(a method of
packet routing)
Avoiding flow-control and error checking at the intermediate
nodes
Small and fixed length unit of data transmitted(53bytes= header
5 + body 48)
 reduction of buffer size, complexity, queuing delay at intermediate nodes
B-ISDN(CCITT I.150 standard)
Optical fiber transmission medium(155 - 622 megabits/sec)
ATM protocol layers
83
Asynchronous Transfer Mode(ATM)- ctd
 ATM protocol Layer
Me ss ag e
L ay ers
A p plic ation
High er-lay er pro toc ols
A TM ad ap tion la ye r
A TM ce lls
A TM la ye r
A TM virtu al ch an ne ls
Phy sica l
 ATM cell layout
Header: 5 byt es
Virt ual pat h id
Virt ual c hannel id
Flags
Data
53 bytes
84
Operating System Support
 Functionality of the Operating System (OS)
resource management (CPU, memory, …)
 Processes and Threads
Similarities V.S. differences
multi-threaded servers and clients
 Implementation of...
communication primitives
Invocations
85
Operating System Support
 Resource sharing
CPU (single/multiprocessor machines)
 concurrent processes/threads
 communication/synchronization primitives
 process scheduling
memory (static/dynamic allocation to programs)
 memory manager
file storage and devices
 file manager, printer driver, etc
 OS kernel
implements CPU and memory sharing
abstracts hardware
Monolithic VS Microkernel
86
Distributed System layers
Applications, services
Middleware
OS: kernel,
libraries &
servers
OS1
Processes, threads,
communication, ...
OS2
Processes, threads,
communication, ...
Computer &
network hardware
Computer &
network hardware
Node 1
Platform
Node 2
87
Core OS functionality
Process manager
Communication
manager
Thread manager
Memory manager
Supervisor
88
Core OS components
 Process manager
 creation and operations on processes (= address space+threads)
 Threads manager
 threads creation, synchronization, scheduling
 Communication manager
communication between threads (sockets, semaphores)
 in different processes(concurrency)
 on different computers(parallel)
 Memory manager
 physical (RAM) and virtual (disk) memory
 Supervisor
 hardware abstraction (dispatching of interrupts, exceptions, system call traps)
 control of memory managements and hardware cache
89
Why middleware again...
90
Why middleware again...
91
Why middleware again...
 Network OS
ex) UNIX, Windows NT
network transparent access for remote files (NFS)
no task/process scheduling across different nodes
services
 rlogin, telnet, ftp, WWW
92
Why middleware again...ctd
93
Why middleware again...ctd
 Distributed OS (Mach..)
transparent process scheduling across nodes
load balancing
none in use:
cost of switching OS too high, load balancing not always easy to achieve
94
Why middleware again... ctd
: NOS
: NOS
: DOS
Distributed Operating System Services
95