Slides 04

Distributed Systems
(3rd Edition)
Chapter 04: Communication

Version: February 25, 2017
Communication: Foundations Layered Protocols
Basic networking model
Application protocol
Application 7
Presentation protocol
Presentation 6
Session protocol
Session 5
Transport protocol
Transport 4
Network protocol
Network 3
Data link protocol
Data link 2
Physical protocol
Physical 1
Network
Drawbacks
Focus on message-passing only
Often unneeded or unwanted functionality
Violates access transparency
The OSI reference model 2 / 49

Low-level layers
Recap
Physical layer: contains the specification and implementation of bits, and
their transmission between sender and receiver
Data link layer: prescribes the transmission of a series of bits into a frame
to allow for error and flow control
Network layer: describes how packets in a network of computers are to be
routed.
Observation
For many distributed systems, the lowest-level interface is that of the network
layer.

Transport Layer
Important
The transport layer provides the actual communication facilities for most
distributed systems.
Standard Internet protocols

TCP: connection-oriented, reliable, stream-oriented communication
UDP: unreliable (best-effort) datagram communication

Middleware layer
Observation
Middleware is invented to provide common services and protocols that can be
used by many different applications
A rich set of communication protocols
(Un)marshaling of data, necessary for integrated systems
Naming protocols, to allow easy sharing of resources
Security protocols for secure communication
Scaling mechanisms, such as for replication and caching
Note
What remains are truly application-specific protocols... such as?
Middleware protocols 5 / 49
An adapted layering scheme
Application protocol
Application
Middleware protocol
Middleware
Operating Host-to-host protocol

system
Physical/Link-level protocol
Hardware
Network
Middleware protocols 6 / 49
Communication: Foundations Types of Communication
Types of communication
Distinguish...
Synchronize at Synchronize at Synchronize after
request submission request delivery processing by server
Client
Request
Transmission
interrupt
Storage
facility
Reply
Server Time
Transient versus persistent communication

Asynchronous versus synchronous communication
7 / 49
Transient versus persistent
Client
Request
Transmission
interrupt
Storage
facility
Reply
Server Time
Transient communication: Comm. server discards message when it

cannot be delivered at the next server, or at the receiver.
Persistent communication: A message is stored at a communication
server as long as it takes to deliver it.
8 / 49
Places for synchronization

Client
Request
Transmission
interrupt
Storage
facility
Reply
Server Time
At request submission
At request delivery
After request processing
9 / 49
Client/Server
Some observations
Client/Server computing is generally based on a model of transient
synchronous communication:
Client and server have to be active at time of communication
Client issues request and blocks until it receives reply
Server essentially waits only for incoming requests, and subsequently
processes them
10 / 49
Client/Server
Some observations
Client/Server computing is generally based on a model of transient
synchronous communication:
Client and server have to be active at time of communication
Client issues request and blocks until it receives reply
Server essentially waits only for incoming requests, and subsequently
processes them
Drawbacks synchronous communication

Client cannot do any other work while waiting for reply
Failures have to be handled immediately: the client is waiting
The model may simply not be appropriate (mail, news)
10 / 49
Messaging
Message-oriented middleware
Aims at high-level persistent asynchronous communication:
Processes send each other messages, which are queued
Sender need not wait for immediate reply, but can do other things
Middleware often ensures fault tolerance
11 / 49
Communication: Remote procedure call Basic RPC operation
Basic RPC operation
Observations
Application developers are familiar with simple procedure model
Well-engineered procedures operate in isolation (black box)
There is no fundamental reason not to execute procedures on separate
machine
Wait for result
Client
Conclusion Call remote Return

procedure from call
Communication between caller & callee
can be hidden by using procedure-call Request Reply
mechanism. Server
Call local procedure Time
and return results
12 / 49
Communication: Remote procedure call Basic RPC operation
Basic RPC operation

Client machine Server machine
Client process Server process

1. Client call to
procedure Implementation 6. Stub makes
of doit local call to “doit”
Server stub
r = doit(a,b) r = doit(a,b)
Client stub
proc: “doit” proc: “doit”
type1: val(a) type1: val(a) 5. Stub unpacks
2. Stub builds message
type2: val(b) message type2: val(b)
proc: “doit” 4. Server OS

Client OS type1: val(a) Server OS hands message
type2: val(b) to server stub
3. Message is sent
across the network
1 Client procedure calls client stub. 6 Server does local call; returns result to stub.
2 Stub builds message; calls local OS. 7 Stub builds message; calls OS.
3 OS sends message to remote OS. 8 OS sends message to client’s OS.
4 Remote OS gives message to stub. 9 Client’s OS gives message to stub.
5 Stub unpacks parameters; calls 10 Client stub unpacks result; returns to client.
server.
13 / 49
Communication: Remote procedure call Parameter passing
RPC: Parameter passing
There’s more than just wrapping parameters into a message

Client and server machines may have different data representations (think
of byte ordering)
Wrapping a parameter means transforming a value into a sequence of
bytes
Client and server have to agree on the same encoding:
How are basic data values represented (integers, floats, characters)

How are complex data values represented (arrays, unions)
Conclusion
Client and server need to properly interpret messages, transforming them into
machine-dependent representations.
14 / 49
Some assumptions
Copy in/copy out semantics: while procedure is executed, nothing can be
assumed about parameter values.
All data that is to be operated on is passed by parameters. Excludes
passing references to (global) data.
15 / 49
Some assumptions
Conclusion
Full access transparency cannot be realized.
15 / 49
Some assumptions
Conclusion
Full access transparency cannot be realized.
A remote reference mechanism enhances access transparency

Remote reference offers unified access to remote data
Remote references can be passed as parameter in RPCs
Note: stubs can sometimes be used as such references
15 / 49
Communication: Remote procedure call Variations on RPC
Asynchronous RPCs
Essence
Try to get rid of the strict request-reply behavior, but let the client continue
without waiting for an answer from the server.
Wait for Callback to client

Client acceptance
Call remote Return

procedure from call Return
results
Accept
Request request
Server Call local procedure Time
Asynchronous RPC 16 / 49
Communication: Remote procedure call Variations on RPC
Sending out multiple RPCs
Essence
Sending an RPC request to a group of servers.
Server Call local procedure
Callbacks to client
Client
Call remote
procedures
Server Call local procedure Time
Multicast RPC 17 / 49
Communication: Remote procedure call Example: DCE RPC
RPC in practice
Uuidgen
Interface
definition file
IDL compiler
Client code Client stub Header Server stub Server code
#include #include
C compiler C compiler C compiler C compiler
Client Client stub Server stub Server

object file object file object file object file
Runtime Runtime
Linker Linker
library library
Client Server
binary binary
Writing a Client and a Server 18 / 49

Communication: Remote procedure call Example: DCE RPC
Client-to-server binding (DCE)
Issues
(1) Client must locate server machine, and (2) locate the server.
Directory machine
Directory
server
2. Register service
3. Look up server
Server machine
Client machine
5. Do RPC 1. Register port

Server
Client
4. Ask for port DCE

daemon Port
table
Binding a client to a server 19 / 49

Communication: Message-oriented communication Simple transient messaging with sockets
Transient messaging: sockets

Berkeley socket interface
Operation Description
socket Create a new communication end point
bind Attach a local address to a socket
listen Tell operating system what the maximum number of pending
connection requests should be
accept Block caller until a connection request arrives
connect Actively attempt to establish a connection
send Send some data over the connection
receive Receive some data over the connection
close Release the connection
Server
socket bind listen accept receive send close
Synchronization point Communication
socket connect send receive close

Client
20 / 49
Communication: Message-oriented communication Simple transient messaging with sockets
Sockets: Python code

Server
1 from socket import *
2 s = socket(AF_INET, SOCK_STREAM)
3 s.bind((HOST, PORT))
4 s.listen(1)
5 (conn, addr) = s.accept() # returns new socket and addr. client
6 while True: # forever
7 data = conn.recv(1024) # receive data from client
8 if not data: break # stop if client stopped
9 conn.send(str(data)+"*") # return sent data plus an "*"
10 conn.close() # close the connection
Client
1 from socket import *
2 s = socket(AF_INET, SOCK_STREAM)
3 s.connect((HOST, PORT)) # connect to server (block until accepted)
4 s.send(’Hello, world’) # send same data
5 data = s.recv(1024) # receive the response
6 print data # print the result
7 s.close() # close the connection
21 / 49
Communication: Message-oriented communication Advanced transient messaging
Making sockets easier to work with
Observation
Sockets are rather low level and programming mistakes are easily made.
However, the way that they are used is often the same (such as in a
client-server setting).
Alternative: ZeroMQ
Provides a higher level of expression by pairing sockets: one for sending
messages at process P and a corresponding one at process Q for receiving
messages. All communication is asynchronous.
Three patterns
Request-reply
Publish-subscribe
Pipeline
Using messaging patterns: ZeroMQ 22 / 49

Request-reply
Server
1 import zmq
2 context = zmq.Context()
3
4 p1 = "tcp://"+ HOST +":"+ PORT1 # how and where to connect
5 p2 = "tcp://"+ HOST +":"+ PORT2 # how and where to connect
6 s = context.socket(zmq.REP) # create reply socket
7
8 s.bind(p1) # bind socket to address
9 s.bind(p2) # bind socket to address
10 while True:
11 message = s.recv() # wait for incoming message
12 if not "STOP" in message: # if not to stop...
13 s.send(message + "*") # append "*" to message
14 else: # else...
15 break # break out of loop and end

Request-reply
Client
1 import zmq
3
4 php = "tcp://"+ HOST +":"+ PORT # how and where to connect
5 s = context.socket(zmq.REQ) # create socket
6
7 s.connect(php) # block until connected
8 s.send("Hello World") # send message
9 message = s.recv() # block until response
10 s.send("STOP") # tell server to stop
11 print message # print result

Publish-subscribe
Server
1 import zmq, time
2
4 s = context.socket(zmq.PUB) # create a publisher socket
5 p = "tcp://"+ HOST +":"+ PORT # how and where to communicate
6 s.bind(p) # bind socket to the address
7 while True:
8 time.sleep(5) # wait every 5 seconds
9 s.send("TIME " + time.asctime()) # publish the current time
Client
1 import zmq
2
4 s = context.socket(zmq.SUB) # create a subscriber socket
5 p = "tcp://"+ HOST +":"+ PORT # how and where to communicate
6 s.connect(p) # connect to the server
7 s.setsockopt(zmq.SUBSCRIBE, "TIME") # subscribe to TIME messages
8
9 for i in range(5): # Five iterations
10 time = s.recv() # receive a message
11 print time
Pipeline
Source
1 import zmq, time, pickle, sys, random
2
4 me = str(sys.argv[1])
5 s = context.socket(zmq.PUSH) # create a push socket
6 src = SRC1 if me == ’1’ else SRC2 # check task source host
7 prt = PORT1 if me == ’1’ else PORT2 # check task source port
8 p = "tcp://"+ src +":"+ prt # how and where to connect
9 s.bind(p) # bind socket to address
10
11 for i in range(100): # generate 100 workloads
12 workload = random.randint(1, 100) # compute workload
13 s.send(pickle.dumps((me,workload))) # send workload to worker

Pipeline
Worker
1 import zmq, time, pickle, sys
2
4 me = str(sys.argv[1])
5 r = context.socket(zmq.PULL) # create a pull socket
6 p1 = "tcp://"+ SRC1 +":"+ PORT1 # address first task source
7 p2 = "tcp://"+ SRC2 +":"+ PORT2 # address second task source
8 r.connect(p1) # connect to task source 1
9 r.connect(p2) # connect to task source 2
10
11 while True:
12 work = pickle.loads(r.recv()) # receive work from a source
13 time.sleep(work[1]*0.01) # pretend to work

MPI: When lots of flexibility is needed
Representative operations
MPI bsend Append outgoing message to a local send buffer
MPI send Send a message and wait until copied to local or
remote buffer
MPI ssend Send a message and wait until transmission starts
MPI sendrecv Send a message and wait for reply
MPI isend Pass reference to outgoing message, and continue
MPI issend Pass reference to outgoing message, and wait until
receipt starts
MPI recv Receive a message; block if there is none
MPI irecv Check if there is an incoming message, but do not
block
The Message-Passing Interface (MPI) 28 / 49

Communication: Message-oriented communication Message-oriented persistent communication
Message-oriented middleware
Essence
Asynchronous persistent communication through support of middleware-level
queues. Queues correspond to buffers at communication servers.
Operations
put Append a message to a specified queue
get Block until the specified queue is nonempty, and
remove the first message
poll Check a specified queue for messages, and remove
the first. Never block
notify Install a handler to be called when a message is put
into the specified queue
Message-queuing model 29 / 49
General model
Queue managers
Queues are managed by queue managers. An application can put messages
only into a local queue. Getting a message is possible by extracting it from a
local queue only ⇒ queue managers need to route messages.
Routing
Look up
Source queue contact address Destination queue
manager of destination manager
queue manager
Logical
queue-level
address (name)
Local OS Address lookup

Local OS
database
Contact
Network address
General architecture of a message-queuing system 30 / 49

Message broker
Observation
Message queuing systems assume a common messaging protocol: all
applications agree on message format (i.e., structure and data representation)
Broker handles application heterogeneity in an MQ system

Transforms incoming messages to target format
Very often acts as an application gateway
May provide subject-based routing capabilities (i.e., publish-subscribe
capabilities)
Message brokers 31 / 49
Message broker: general architecture
Source Message broker Destination
Application Application
Broker plugins Rules
Interface Interface
Queuing
layer
Local OS Local OS Local OS
Message brokers 32 / 49
Communication: Message-oriented communication Example: IBM’s WebSphere message-queuing system
IBM’s WebSphere MQ
Basic concepts
Application-specific messages are put into, and removed from queues
Queues reside under the regime of a queue manager
Processes can put messages only in local queues, or through an RPC
mechanism
Message transfer
Messages are transferred between queues
Message transfer between queues at different processes, requires a
channel
At each end point of channel is a message channel agent
Message channel agents are responsible for:
Setting up channels using lower-level network communication
facilities (e.g., TCP/IP)
(Un)wrapping messages from/in transport-level packets
Sending/receiving packets
Overview 33 / 49
Schematic overview
Client's receive
Sending client Routing table Send queue queue Receiving client
Queue Queue
manager manager
MQ Interface MQ Interface
Server MCA MCA Server

Stub MCA MCA Stub
stub stub
Enterprise network Local network

RPC Message passing To other remote
(synchronous) (asynchronous) queue managers
Channels are inherently unidirectional

Automatically start MCAs when messages arrive
Any network of queue managers can be created
Routes are set up manually (system administration)
Overview 34 / 49
Message channel agents
Some attributes associated with message channel agents

Attribute Description
Transport type Determines the transport protocol to be used
FIFO delivery Indicates that messages are to be delivered in the
order they are sent
Message length Maximum length of a single message
Setup retry count Specifies maximum number of retries to start up the
remote MCA
Delivery retries Maximum times MCA will try to put received message
into queue
Channels 35 / 49
Routing
By using logical names, in combination with name resolution to local queues, it
is possible to put a message in a remote queue
Alias table Routing table

LA1 QMC QMB SQ1 Routing table
Alias table
LA2 QMD QMC SQ1
LA1 QMA QMA SQ1
QMD SQ2
LA2 QMD QMC SQ1
QMD SQ1
SQ2 SQ1
SQ1
QMA
QMB
Routing table Routing table

SQ1 QMB
QMA SQ1 QMA SQ1
QMC SQ2 QMC SQ1
QMB SQ1 SQ2 QMD SQ1
Alias table
LA1 QMA
SQ1
LA2 QMC
QMD
Message transfer 36 / 49
Communication: Multicast communication Application-level tree-based multicasting
Application-level multicasting
Essence
Organize nodes of a distributed system into an overlay network and use that
network to disseminate data:
Oftentimes a tree, leading to unique paths
Alternatively, also mesh networks, requiring a form of routing
37 / 49
Application-level multicasting in Chord
Basic approach
1 Initiator generates a multicast identifier mid.
2 Lookup succ(mid), the node responsible for mid.
3 Request is routed to succ(mid), which will become the root.
4 If P wants to join, it sends a join request to the root.
5 When request arrives at Q:
Q has not seen a join request before ⇒ it becomes forwarder; P
becomes child of Q. Join request continues to be forwarded.
Q knows about tree ⇒ P becomes child of Q. No need to forward
join request anymore.
38 / 49
ALM: Some costs
Different metrics
End host E Router
A 1 1 1 C
Ra Re Rc
30 20
5
7
40 Rd
1 Rb 1
Internet D
B
Overlay network
Link stress: How often does an ALM message cross the same physical
link? Example: message from A to D needs to cross hRa, Rbi twice.
Stretch: Ratio in delay between ALM-level path and network-level path.
Example: messages B to C follow path of length 73 at ALM, but 47 at
network level ⇒ stretch = 73/47.
Performance issues in overlays 39 / 49

Communication: Multicast communication Flooding-based multicasting
Flooding
Essence The size of a random overlay as

P simply sends a message m to function of the number of nodes
each of its neighbors. Each 300
Number of edges (x 1000)

neighbor will forward that message, 250
except to P, and only if it had not 200 pedge = 0.6
seen m before. 150
pedge = 0.4
100
Performance 50
pedge = 0.2
The more edges, the more 0

expensive! 100 500 1000
Number of nodes
40 / 49
Communication: Multicast communication Flooding-based multicasting
Flooding
Essence The size of a random overlay as

P simply sends a message m to function of the number of nodes
each of its neighbors. Each 300
Number of edges (x 1000)

neighbor will forward that message, 250
except to P, and only if it had not 200 pedge = 0.6
seen m before. 150
pedge = 0.4
100
Performance 50
pedge = 0.2
The more edges, the more 0

expensive! 100 500 1000
Number of nodes
Variation
Let Q forward a message with a certain probability pflood , possibly even
dependent on its own number of neighbors (i.e., node degree) or the degree of
its neighbors.
40 / 49
Communication: Multicast communication Gossip-based data dissemination
Epidemic protocols
Assume there are no write–write conflicts

Update operations are performed at a single server
A replica passes updated state to only a few neighbors
Update propagation is lazy, i.e., not immediate
Eventually, each update should reach every replica
Two forms of epidemics

Anti-entropy: Each replica regularly chooses another replica at random,
and exchanges state differences, leading to identical states at both
afterwards
Rumor spreading: A replica which has just been updated (i.e., has been
contaminated), tells a number of other replicas about its update
(contaminating them as well).
41 / 49
Anti-entropy
Principle operations
A node P selects another node Q from the system at random.
Pull: P only pulls in new updates from Q
Push: P only pushes its own updates to Q
Push-pull: P and Q send updates to each other
Observation
For push-pull it takes O(log(N)) rounds to disseminate updates to all N nodes
(round = when every node has taken the initiative to start an exchange).
Information dissemination models 42 / 49

Anti-entropy: analysis
Basics
Consider a single source, propagating its update. Let pi be the probability that
a node has not received the update after the i th round.
Analysis: staying ignorant
With pull, pi +1 = (pi )2 : the node was

not updated during the i th round and 1.0
N = 10,000
Probability not yet updated

should contact another ignorant node 0.8
during the next round.
push
With push, 0.6
pi +1 = pi (1 − N1 )N (1−pi ) ≈ pi e−1 (for 0.4 push-pull

small pi and large N): the node was pull
ignorant during the i th round and no 0.2
updated node chooses to contact it

during the next round. 0 5 10 15 20 25
Round
With push-pull: (pi )2 · (pi e−1 )

Anti-entropy performance
1.0
N = 10,000
Probability not yet updated 0.8
0.6 push
0.4 push-pull
pull
0.2
0 5 10 15 20 25
Round

Rumor spreading
Basic model
A server S having an update to report, contacts other servers. If a server is
contacted to which the update has already propagated, S stops contacting
other servers with probability pstop .
Observation
If s is the fraction of ignorant servers (i.e., which are unaware of the update), it
can be shown that with many servers
−(1/pstop +1)(1−s)
s=e

Formal analysis
Notations
Let s denote fraction of nodes that have not yet been updated (i.e., susceptible;
i the fraction of updated (infected) and active nodes; and r the fraction of
updated nodes that gave up (removed).
From theory of epidemics

(1) ds/dt = −s · i
(2) di/dt = s · i − pstop · (1 − s) · i
p
⇒ di/ds = −(1 + pstop ) + stop
s
⇒ i(s) = −(1 + pstop ) · s + pstop · ln(s) + C
Wrapup
i(1) = 0 ⇒ C = 1 + pstop ⇒ i(s) = (1 + pstop ) · (1 − s) + pstop · ln(s). We are
−(1/pstop +1)(1−s)
looking for the case i(s) = 0, which leads to s = e

Rumor spreading
The effect of stopping
0.20 Consider 10,000 nodes

1/pstop s Ns
0.15
1 0.203188 2032
s 0.10 2 0.059520 595
3 0.019827 198
0.05
4 0.006977 70
0.00 5 0.002516 25
0.0 0.2 0.4 0.6 0.8 1.0 6 0.000918 9
pstop
7 0.000336 3

Rumor spreading
The effect of stopping
0.20 Consider 10,000 nodes

1/pstop s Ns
0.15
1 0.203188 2032
s 0.10 2 0.059520 595
3 0.019827 198
0.05
4 0.006977 70
0.00 5 0.002516 25
0.0 0.2 0.4 0.6 0.8 1.0 6 0.000918 9
pstop
7 0.000336 3
Note
If we really have to ensure that all servers are eventually updated, rumor
spreading alone is not enough

Deleting values
Fundamental problem
We cannot remove an old value from a server and expect the removal to
propagate. Instead, mere removal will be undone in due time using epidemic
algorithms
Solution
Removal has to be registered as a special update by inserting a death
certificate
Removing data 48 / 49
Deleting values
When to remove a death certificate (it is not allowed to stay for ever)
Run a global algorithm to detect whether the removal is known
everywhere, and then collect the death certificates (looks like garbage
collection)
Assume death certificates propagate in finite time, and associate a
maximum lifetime for a certificate (can be done at risk of not reaching all
servers)
Note
It is necessary that a removal actually reaches all servers.
Removing data 49 / 49

Slides 04

Uploaded by

Copyright:

Available Formats

Slides 04

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Slides 04

Uploaded by

Copyright:

Available Formats

Distributed Systems

Chapter 04: Communication

Basic networking model

The OSI reference model 2 / 49

The OSI reference model 3 / 49

Standard Internet protocols

The OSI reference model 4 / 49

An adapted layering scheme

Operating Host-to-host protocol

Transient versus persistent communication

Transient communication: Comm. server discards message when it

Places for synchronization

Drawbacks synchronous communication

Basic RPC operation

Conclusion Call remote Return

Basic RPC operation

Client process Server process

proc: “doit” 4. Server OS

RPC: Parameter passing

There’s more than just wrapping parameters into a message

How are basic data values represented (integers, floats, characters)

RPC: Parameter passing

RPC: Parameter passing

RPC: Parameter passing

A remote reference mechanism enhances access transparency

Wait for Callback to client

Call remote Return

Server Call local procedure Time

Sending out multiple RPCs

Server Call local procedure

Server Call local procedure Time

Client code Client stub Header Server stub Server code

C compiler C compiler C compiler C compiler

Client Client stub Server stub Server

Writing a Client and a Server 18 / 49

Client-to-server binding (DCE)

5. Do RPC 1. Register port

4. Ask for port DCE

Binding a client to a server 19 / 49

Transient messaging: sockets

Synchronization point Communication

socket connect send receive close

Sockets: Python code

Making sockets easier to work with

Using messaging patterns: ZeroMQ 22 / 49

Using messaging patterns: ZeroMQ 23 / 49

Using messaging patterns: ZeroMQ 24 / 49

Using messaging patterns: ZeroMQ 26 / 49

Using messaging patterns: ZeroMQ 27 / 49

MPI: When lots of flexibility is needed

The Message-Passing Interface (MPI) 28 / 49

Local OS Address lookup

General architecture of a message-queuing system 30 / 49

Broker handles application heterogeneity in an MQ system

Message broker: general architecture

Source Message broker Destination

Broker plugins Rules

Server MCA MCA Server