Carnegie Mellon
Introduction to Computer Systems15-213/18-243, spring 200922nd Lecture, Apr. 9th
Instructors: Gregory Kesden and Markus Püschel
Carnegie Mellon
Exam 2 (This Section)
Carnegie Mellon
Last Time: Open Files in Unix Two descriptors referencing two distinct open disk files.
Descriptor 1 (stdout) points to terminal, and descriptor 4 points to open disk file
fd 0fd 1fd 2fd 3fd 4
Descriptor table[one table per process]
Open file table [shared by all processes]
v-node table[shared by all processes]
File posrefcnt=1
...
File posrefcnt=1
...
stderrstdoutstdin File access
...
File sizeFile type
File access
...
File sizeFile type
File A (terminal)
File B (disk)
Info in stat struct
Carnegie Mellon
Last Time: Unix I/O
file fd
memory
pos pos (after read or write)
n bytes
buf
read write
ssize_t read(int fd, void *buf, size_t n)ssize_t write(int fd, void *buf, size_t n)
Carnegie Mellon
Last Time: Standard I/O (Buffering)
memory
fread fwrite
file fd
buffer
data transferred in chunks
abstracted as stream:
FILE
Carnegie Mellon
Last Time: Robust I/O rio_readn/* * rio_readn - robustly read n bytes (unbuffered) */ssize_t rio_readn(int fd, void *usrbuf, size_t n) { size_t nleft = n; ssize_t nread; char *bufp = usrbuf;
while (nleft > 0) {if ((nread = read(fd, bufp, nleft)) < 0) { if (errno == EINTR) /* interrupted by sig handler return */
nread = 0; /* and call read() again */ else
return -1; /* errno set by read() */ } else if (nread == 0) break; /* EOF */nleft -= nread;bufp += nread;
} return (n - nleft); /* return >= 0 */}
Carnegie Mellon
Today System level I/O
Unix I/O Standard I/O RIO (robust I/O) package Conclusions and examples
Internetworking Networks Global IP Internet
Carnegie Mellon
Buffered I/O: Motivation I/O Applications Read/Write One Character at a Time
getc, putc, ungetc gets
Read line of text, stopping at newline Implementing as Calls to Unix I/O Expensive
Read & Write involve require Unix kernel calls > 10,000 clock cycles
Buffered Read Use Unix read() to grab block of bytes User input functions take one byte at a time from buffer
Refill buffer when empty
unreadalready readBuffer
Carnegie Mellon
unread
Buffered I/O: Implementation For reading from file File has associated buffer to hold bytes that have been read
from file but not yet read by user code
Layered on Unix File
already readBuffer
rio_bufrio_bufptr
rio_cnt
unreadalready readnot in buffer unseen
Current File Position
Buffered Portion
Carnegie Mellon
Buffered I/O: Declaration All information contained in struct
typedef struct { int rio_fd; /* descriptor for this internal buf */ int rio_cnt; /* unread bytes in internal buf */ char *rio_bufptr; /* next unread byte in internal buf */ char rio_buf[RIO_BUFSIZE]; /* internal buffer */} rio_t;
unreadalready readBuffer
rio_bufrio_bufptr
rio_cnt
Carnegie Mellon
RIO: Robustness and Buffering
Robust
Buffered
read
rio_read
rio_readn
rio_readnb
Carnegie Mellon
Buffered RIO Input Functions Efficiently read text lines and binary data from a file partially
cached in an internal memory buffer
rio_readlineb reads a text line of up to maxlen bytes from file fd and stores the line in usrbuf
Especially useful for reading text lines from network sockets Stopping conditions
maxlen bytes read EOF encountered Newline (‘\n’) encountered
#include "csapp.h"
void rio_readinitb(rio_t *rp, int fd);
ssize_t rio_readlineb(rio_t *rp, void *usrbuf, size_t maxlen);
Return: num. bytes read if OK, 0 on EOF, -1 on error
Carnegie Mellon
Buffered RIO Input Functions (cont)
rio_readnb reads up to n bytes from file fd Stopping conditions
maxlen bytes read EOF encountered
Calls to rio_readlineb and rio_readnb can be interleaved arbitrarily on the same descriptor
Warning: Don’t interleave with calls to rio_readn
#include "csapp.h"
void rio_readinitb(rio_t *rp, int fd);
ssize_t rio_readlineb(rio_t *rp, void *usrbuf, size_t maxlen);ssize_t rio_readnb(rio_t *rp, void *usrbuf, size_t n);
Return: num. bytes read if OK, 0 on EOF, -1 on error
Carnegie Mellon
RIO Example Copying the lines of a text file from standard input to
standard output
#include "csapp.h"
int main(int argc, char **argv) { int n; rio_t rio; char buf[MAXLINE];
Rio_readinitb(&rio, STDIN_FILENO); while((n = Rio_readlineb(&rio, buf, MAXLINE)) != 0)
Rio_writen(STDOUT_FILENO, buf, n); exit(0);}
Carnegie Mellon
Today System level I/O
Unix I/O Standard I/O RIO (robust I/O) package Conclusions and examples
Internetworking Networks Global IP Internet
Carnegie Mellon
Choosing I/O Functions General rule: use the highest-level I/O functions you can
Many C programmers are able to do all of their work using the standard I/O functions
When to use standard I/O When working with disk or terminal files
When to use raw Unix I/O When you need to fetch file metadata In rare cases when you need absolute highest performance
When to use RIO When you are reading and writing network sockets or pipes Never use standard I/O or raw Unix I/O on sockets or pipes
Carnegie Mellon
For Further Information The Unix bible:
W. Richard Stevens & Stephen A. Rago, Advanced Programming in the Unix Environment, 2nd Edition, Addison Wesley, 2005
Updated from Stevens’ 1993 book
Stevens is arguably the best technical writer ever. Produced authoritative works in:
Unix programming TCP/IP (the protocol that makes the Internet work) Unix network programming Unix IPC programming
Tragically, Stevens died Sept. 1, 1999 But others have taken up his legacy
Carnegie Mellon
Fun with File Descriptors (1)
What would this program print for file containing “abcde”?
#include "csapp.h"int main(int argc, char *argv[]){ int fd1, fd2, fd3; char c1, c2, c3; char *fname = argv[1]; fd1 = Open(fname, O_RDONLY, 0); fd2 = Open(fname, O_RDONLY, 0); fd3 = Open(fname, O_RDONLY, 0); Dup2(fd2, fd3); Read(fd1, &c1, 1); Read(fd2, &c2, 1); Read(fd3, &c3, 1); printf("c1 = %c, c2 = %c, c3 = %c\n", c1, c2, c3); return 0;}
Carnegie Mellon
Fun with File Descriptors (2)
What would this program print for file containing “abcde”?
#include "csapp.h"int main(int argc, char *argv[]){ int fd1; int s = getpid() & 0x1; char c1, c2; char *fname = argv[1]; fd1 = Open(fname, O_RDONLY, 0); Read(fd1, &c1, 1); if (fork()) { /* Parent */ sleep(s); Read(fd1, &c2, 1); printf("Parent: c1 = %c, c2 = %c\n", c1, c2); } else { /* Child */ sleep(1-s); Read(fd1, &c2, 1); printf("Child: c1 = %c, c2 = %c\n", c1, c2); } return 0;}
Carnegie Mellon
Fun with File Descriptors (3)
What would be the contents of the resulting file?
#include "csapp.h"int main(int argc, char *argv[]){ int fd1, fd2, fd3; char *fname = argv[1]; fd1 = Open(fname, O_CREAT|O_TRUNC|O_RDWR, S_IRUSR|S_IWUSR); Write(fd1, "pqrs", 4); fd3 = Open(fname, O_APPEND|O_WRONLY, 0); Write(fd3, "jklmn", 5); fd2 = dup(fd1); /* Allocates descriptor */ Write(fd2, "wxyz", 4); Write(fd3, "ef", 2); return 0;}
Carnegie Mellon
Accessing Directories Only recommended operation on a directory: read its entries
dirent structure contains information about a directory entry DIR structure contains information about directory while stepping
through its entries
#include <sys/types.h>#include <dirent.h>
{ DIR *directory; struct dirent *de; ... if (!(directory = opendir(dir_name))) error("Failed to open directory"); ... while (0 != (de = readdir(directory))) { printf("Found file: %s\n", de->d_name); } ... closedir(directory);}
Carnegie Mellon
Unix I/O vs. Standard I/O vs. RIO Standard I/O and RIO are implemented using low-level
Unix I/O
Which ones should you use in your programs?
Unix I/O functions (accessed via system calls)
Standard I/O functions
C application program
fopen fdopenfread fwrite fscanf fprintf sscanf sprintf fgets fputs fflush fseekfclose
open readwrite lseekstat close
rio_readnrio_writenrio_readinitbrio_readlinebrio_readnb
RIOfunctions
Carnegie Mellon
Pros and Cons of Unix I/O Pros
Unix I/O is the most general and lowest overhead form of I/O All other I/O packages are implemented using Unix I/O
functions Unix I/O provides functions for accessing file metadata
Cons Dealing with short counts is tricky and error prone Efficient reading of text lines requires some form of buffering, also
tricky and error prone Both of these issues are addressed by the standard I/O and RIO
packages
Carnegie Mellon
Pros and Cons of Standard I/O Pros
Buffering increases efficiency by decreasing the number of read and write system calls
Short counts are handled automatically
Cons Provides no function for accessing file metadata Standard I/O is not appropriate for input and output on network
sockets There are poorly documented restrictions on streams that interact
badly with restrictions on sockets
Carnegie Mellon
Today System level I/O
Unix I/O Standard I/O RIO (robust I/O) package Conclusions and examples
Internetworking Networks Global IP Internet
Carnegie Mellon
A Client-Server Transaction
Clientprocess
Serverprocess
1. Client sends request
2. Server handlesrequest
3. Server sends response4. Client handles
response
Resource
Most network applications are based on the client-server model: A server process and one or more client processes Server manages some resource Server provides service by manipulating resource for clients Server activated by request from client (vending machine analogy)
Note: clients and servers are processes running on hosts (can be the same or different hosts)
Carnegie Mellon
Hardware Organization of a Network Host
mainmemory
I/O bridgeMI
ALU
register fileCPU chip
system bus memory bus
disk controller
graphicsadapter
USBcontroller
mouse keyboard monitordisk
I/O bus
Expansion slots
networkadapter
network
Carnegie Mellon
Computer Networks A network is a hierarchical system of boxes and wires
organized by geographical proximity SAN (System Area Network) spans cluster or machine room
Switched Ethernet, Quadrics QSW, … LAN (Local Area Network) spans a building or campus
Ethernet is most prominent example WAN (Wide Area Network) spans country or world
Typically high-speed point-to-point phone lines
An internetwork (internet) is an interconnected set of networks The Global IP Internet (uppercase “I”) is the most famous example
of an internet (lowercase “i”)
Let’s see how an internet is built from the ground up
Carnegie Mellon
Lowest Level: Ethernet Segment
Ethernet segment consists of a collection of hosts connected by wires (twisted pairs) to a hub
Spans room or floor in a building Operation
Each Ethernet adapter has a unique 48-bit address (MAC address) Hosts send bits to any other host in chunks called frames Hub slavishly copies each bit from each port to every other port
Every host sees every bit Note: Hubs are on their way out. Bridges (switches, routers) became cheap enough
to replace them (means no more broadcasting)
host host host
hub100 Mb/s100 Mb/s
port
Carnegie Mellon
Next Level: Bridged Ethernet Segment
Spans building or campus Bridges cleverly learn which hosts are reachable from which
ports and then selectively copy frames from port to port
host host host host host
hub
hub
bridge
100 Mb/s 100 Mb/s
host host
hub
100 Mb/s 100 Mb/s
1 Gb/s
host host host
bridge
hosthost
hub
A B
C
X
Y
Carnegie Mellon
Conceptual View of LANs For simplicity, hubs, bridges, and wires are often shown as a
collection of hosts attached to a single wire:
host host host...
Carnegie Mellon
Next Level: internets Multiple incompatible LANs can be physically connected by
specialized computers called routers The connected networks are called an internet
host host host... host host host...
WAN WAN
LAN 1 and LAN 2 might be completely different, totally incompatible (e.g., Ethernet and Wifi, 802.11*, T1-links, DSL, …)
router router routerLAN LAN
Carnegie Mellon
Logical Structure of an internet
Ad hoc interconnection of networks No particular topology Vastly different router & link capacities
Send packets from source to destination by hopping through networks Router forms bridge from one network to another Different packets may take different routes
router
router
routerrouter
router
router
hosthost
Carnegie Mellon
The Notion of an internet Protocol How is it possible to send bits across incompatible LANs
and WANs?
Solution: protocol software running on each host and router smooths out the differences between the different networks
Implements an internet protocol (i.e., set of rules) governs how hosts and routers should cooperate when they
transfer data from network to network TCP/IP is the protocol for the global IP Internet
Carnegie Mellon
What Does an internet Protocol Do? Provides a naming scheme
An internet protocol defines a uniform format for host addresses Each host (and router) is assigned at least one of these internet
addresses that uniquely identifies it
Provides a delivery mechanism An internet protocol defines a standard transfer unit (packet) Packet consists of header and payload
Header: contains info such as packet size, source and destination addresses
Payload: contains data bits sent from source host
Carnegie Mellon
LAN2
Transferring Data Over an internet
protocolsoftware
client
LAN1adapter
Host ALAN1
data(1)
data PH FH1(4)
data PH FH2(6)
data(8)
data PH FH2 (5)
LAN2 frame
protocolsoftware
LAN1adapter
LAN2adapter
Routerdata PH(3) FH1
data PH FH1(2)
internet packet
LAN1 frame
(7) data PH FH2
protocolsoftware
server
LAN2adapter
Host B
PH: Internet packet headerFH: LAN frame header
Carnegie Mellon
Other Issues We are glossing over a number of important questions:
What if different networks have different maximum frame sizes? (segmentation)
How do routers know where to forward frames? How are routers informed when the network topology changes? What if packets get lost?
These (and other) questions are addressed by the area of systems known as computer networking
Carnegie Mellon
Today System level I/O
Unix I/O Standard I/O RIO (robust I/O) package Conclusions and examples
Internetworking Networks Global IP Internet
Carnegie Mellon
Global IP Internet Most famous example of an internet
Based on the TCP/IP protocol family IP (Internet protocol) :
Provides basic naming scheme and unreliable delivery capability of packets (datagrams) from host-to-host
UDP (Unreliable Datagram Protocol) Uses IP to provide unreliable datagram delivery from
process-to-process TCP (Transmission Control Protocol)
Uses IP to provide reliable byte streams from process-to-process over connections
Accessed via a mix of Unix file I/O and functions from the sockets interface
Carnegie Mellon
Hardware and Software Organization of an Internet Application
TCP/IP
Client
Networkadapter
Global IP Internet
TCP/IP
Server
Networkadapter
Internet client host Internet server host
Sockets interface(system calls)
Hardware interface(interrupts)
User code
Kernel code
Hardwareand firmware
Carnegie Mellon
Basic Internet Components Internet backbone:
collection of routers (nationwide or worldwide) connected by high-speed point-to-point networks
Network Access Point (NAP): router that connects multiple backbones (often referred to as peers)
Regional networks: smaller backbones that cover smaller geographical areas
(e.g., cities or states) Point of presence (POP):
machine that is connected to the Internet Internet Service Providers (ISPs):
provide dial-up or direct access to POPs
Carnegie Mellon
NAP-Based Internet Architecture NAPs link together commercial backbones provided by
companies such as AT&T and Worldcom
Currently in the US there are about 50 commercial backbones connected by ~12 NAPs (peering points)
Similar architecture worldwide connects national networks to the Internet
Carnegie Mellon
Internet Connection HierarchyNAP NAP
Backbone BackboneBackboneBackbone
NAP
POP POP POP
Regional net
POPPOP POP
POPPOP
Small Business
Big BusinessISP
POP POP POP POP
Pgh employee
Cablemodem
DC employee
POP
T3
T1
ISP (for individuals)
POP
DSLT1
Colocationsites
Private“peering”
agreementsbetween
two backbonecompanies
often bypassNAP
Carnegie Mellon
Network Access Points (NAPs)
Source: Boardwatch.com
Note: Peers in this context are commercial backbones (droh)
Carnegie Mellon
Source: http://personalpages.manchester.ac.uk/staff/m.dodge/cybergeography/atlas/
MCI/WorldCom/UUNET Global Backbone
Carnegie Mellon
Naming and Communicating on the Internet Original Idea
Every node on Internet would have unique IP address Everyone would be able to talk directly to everyone
No secrecy or authentication Messages visible to routers and hosts on same LAN Possible to forge source field in packet header
Shortcomings There aren't enough IP addresses available Don't want everyone to have access or knowledge of all other hosts Security issues mandate secrecy & authentication
Carnegie Mellon
Evolution of Internet: Naming Dynamic address assignment
Most hosts don't need to have known address Only those functioning as servers
DHCP (Dynamic Host Configuration Protocol) Local ISP assigns address for temporary use
Example: My laptop at CMU
IP address 128.2.220.249 (bryant-tp3.cs.cmu.edu) Assigned statically
My laptop at home IP address 205.201.7.7 (dhcp-7-7.dsl.telerama.com) Assigned dynamically by my ISP for my DSL service
Carnegie Mellon
Evolution of Internet: Firewalls
Firewalls Hides organizations nodes from rest of Internet Use local IP addresses within organization For external service, provides proxy service
1. Client request: src=10.2.2.2, dest=216.99.99.992. Firewall forwards: src=176.3.3.3, dest=216.99.99.993. Server responds: src=216.99.99.99, dest=176.3.3.34. Firewall forwards response: src=216.99.99.99, dest=10.2.2.2
Corporation X
Firewall
Internet
10.2.2.214 2
3
176.3.3.3
216.99.99.99
Carnegie Mellon
Virtual Private Networks
Supporting road warrior Employee working remotely with assigned IP address 198.3.3.3 Wants to appear to rest of corporation as if working internally
From address 10.6.6.6 Gives access to internal services (e.g., ability to send mail)
Virtual Private Network (VPN) Overlays private network on top of regular Internet
Corporation X
Internet
10.x.x.x 198.3.3.3Firewall 10.6.6.6
Carnegie Mellon
A Programmer’s View of the Internet Hosts are mapped to a set of 32-bit IP addresses
128.2.203.179
The set of IP addresses is mapped to a set of identifiers called Internet domain names 128.2.203.179 is mapped to www.cs.cmu.edu
A process on one Internet host can communicate with a process on another Internet host over a connection
Carnegie Mellon
IP Addresses 32-bit IP addresses are stored in an IP address struct
IP addresses are always stored in memory in network byte order (big-endian byte order)
True in general for any integer transferred in a packet header from one machine to another.
E.g., the port number used to identify an Internet connection.
/* Internet address structure */struct in_addr { unsigned int s_addr; /* network byte order (big-endian) */};
Useful network byte-order conversion functions:
htonl: convert long int from host to network byte orderhtons: convert short int from host to network byte orderntohl: convert long int from network to host byte orderntohs: convert short int from network to host byte order
Carnegie Mellon
Dotted Decimal Notation By convention, each byte in a 32-bit IP address is represented
by its decimal value and separated by a period IP address: 0x8002C2F2 = 128.2.194.242
Functions for converting between binary IP addresses and dotted decimal strings: inet_aton: dotted decimal string → IP address in network byte order inet_ntoa: IP address in network byte order → dotted decimal string
“n” denotes network representation “a” denotes application representation
Blackboard?
Carnegie Mellon
Dotted Decimal Notation By convention, each byte in a 32-bit IP address is represented
by its decimal value and separated by a period IP address: 0x8002C2F2 = 128.2.194.242
Functions for converting between binary IP addresses and dotted decimal strings: inet_aton: dotted decimal string → IP address in network byte order inet_ntoa: IP address in network byte order → dotted decimal string
“n” denotes network representation “a” denotes application representation
Carnegie Mellon
IP Address Structure IP (V4) Address space divided into classes:
Network ID Written in form w.x.y.z/n n = number of bits in host address E.g., CMU written as 128.2.0.0/16
Class B address Unrouted (private) IP addresses:
10.0.0.0/8 172.16.0.0/12 192.168.0.0/16
Class A
Class B
Class C
Class D
Class E
0 1 2 3 8 16 24 310 Net ID Host ID
Host ID
Host IDNet ID
Net ID
Multicast address
Reserved for experiments
1 01 01
1 1 01
1 1 11
Carnegie Mellon
Internet Domain Names
.net .edu .gov .com
cmu berkeleymit
cs ece
kittyhawk128.2.194.242
cmcl
unnamed root
pdl
imperial128.2.189.40
amazon
www208.216.181.15
First-level domain names
Second-level domain names
Third-level domain names
Carnegie Mellon
Domain Naming System (DNS) The Internet maintains a mapping between IP addresses and
domain names in a huge worldwide distributed database called DNS Conceptually, programmers can view the DNS database as a collection of
millions of host entry structures:
Functions for retrieving host entries from DNS: gethostbyname: query key is a DNS domain name. gethostbyaddr: query key is an IP address.
/* DNS host entry structure */ struct hostent { char *h_name; /* official domain name of host */ char **h_aliases; /* null-terminated array of domain names */ int h_addrtype; /* host address type (AF_INET) */ int h_length; /* length of an address, in bytes */ char **h_addr_list; /* null-terminated array of in_addr structs */ };
Carnegie Mellon
Properties of DNS Host Entries Each host entry is an equivalence class of domain names and
IP addresses Each host has a locally defined domain name localhost
which always maps to the loopback address 127.0.0.1 Different kinds of mappings are possible:
Simple case: one-to-one mapping between domain name and IP address: kittyhawk.cmcl.cs.cmu.edu maps to 128.2.194.242
Multiple domain names mapped to the same IP address: eecs.mit.edu and cs.mit.edu both map to 18.62.1.6
Multiple domain names mapped to multiple IP addresses: aol.com and www.aol.com map to multiple IP addresses
Some valid domain names don’t map to any IP address: for example: cmcl.cs.cmu.edu
Carnegie Mellon
A Program That Queries DNSint main(int argc, char **argv) { /* argv[1] is a domain name */ char **pp; /* or dotted decimal IP addr */ struct in_addr addr; struct hostent *hostp;
if (inet_aton(argv[1], &addr) != 0) hostp = Gethostbyaddr((const char *)&addr, sizeof(addr), AF_INET); else hostp = Gethostbyname(argv[1]); printf("official hostname: %s\n", hostp->h_name); for (pp = hostp->h_aliases; *pp != NULL; pp++) printf("alias: %s\n", *pp);
for (pp = hostp->h_addr_list; *pp != NULL; pp++) { addr.s_addr = ((struct in_addr *)*pp)->s_addr; printf("address: %s\n", inet_ntoa(addr)); }}
Carnegie Mellon
Querying DNS from the Command Line Domain Information Groper (dig) provides a scriptable
command line interface to DNS
linux> dig +short kittyhawk.cmcl.cs.cmu.edu 128.2.194.242 linux> dig +short -x 128.2.194.242 KITTYHAWK.CMCL.CS.CMU.EDU. linux> dig +short aol.com 205.188.145.215 205.188.160.121 64.12.149.24 64.12.187.25 linux> dig +short -x 64.12.187.25 aol-v5.websys.aol.com.
Carnegie Mellon
Internet Connections Clients and servers communicate by sending streams of bytes
over connections: Point-to-point, full-duplex (2-way communication), and reliable.
A socket is an endpoint of a connection Socket address is an IPaddress:port pair
A port is a 16-bit integer that identifies a process: Ephemeral port: Assigned automatically on client when client makes a
connection request Well-known port: Associated with some service provided by a server
(e.g., port 80 is associated with Web servers)
A connection is uniquely identified by the socket addresses of its endpoints (socket pair) (cliaddr:cliport, servaddr:servport)
Carnegie Mellon
Putting it all Together: Anatomy of an Internet Connection
Connection socket pair(128.2.194.242:51213, 208.216.181.15:80)
Server(port 80)Client
Client socket address128.2.194.242:51213
Server socket address208.216.181.15:80
Client host address128.2.194.242
Server host address208.216.181.15