1. Local Area Networks

2. LAN Overview roadmap Why a LAN? How does a computer attach to a LAN?
What’s the difference between switch and a hub?
How is a host configured?
What’s an OS and a NOS?
Network Computing Models: Mainframe, File Server, Client/Server, Peer to Peer.
The LAN is the par t of the network that relates most to the user of the network - its Local to the user, so to speak. We’ll be covering details of Ethernet and TCP/IP in later sessions, so after a brief introduction to the practical details of getting a computer working on a LAN, I took advantage of the broad scope of the subject of LANs to discuss different architectures for network computing and what they’ve evolved from and may be evolving into

3. Why a Local Area Network?
Aggregate users of common resources
Share access to:
Internet
file, database, application or compute servers
printers
Share Resources - Share Costs - Save Money
Share information - Save Time and Increase Productivity

4. Why a Local Area Network?
Workstation
Dedicated 100 Mbps
Shared 10Mbps
Multimedia
Workstation
Ethernet
Hub
Ethernet
Switch
Server
Dedicated 100 Mbps
Router
Printer
Internet or
Corporate Intranet

5. Local Area Networks
“Just as computer networks have grown across continents and oceans to interconnect major computing facilities around the world, they are now growing down corridors and between buildings to interconnect minicomputers in offices and labs.”
Robert M. Metcalfe, Xerox, 1976

6. How does a computer attach to a LAN?
Server
Large disk stores
files and databases,
Fast processor computes
on demand
Personal
Computer
Accesses LAN resources
Wire
Cat 5 UTP
Wiring Hub
Serves as center of network,
contains multiple independent
but connected modules where
network equipment can be connected
Now, in order to have a network, you must have at least two devices that communicate with each other. In this simple model, it is a computer and a printer. The printer also has a NIC installed, which in turn is plugged into a wiring hub. The computer systems is also plugged into the hub, which facilitates communication between the two devices.
Networking Operating
Software/System (NOS)
Coordinates access to LAN,
provides software interface
for PC applications
Network Interface Card (NIC)
Amplifies electronic signals,
packages data for transmission, and
controls access to the network cable

7. How does a computer attach to a LAN?
Computer with appropriate interface bus.
NIC (Network Interface Card)
NOS (Network Operating System)
Cable (normally Cat 5 UTP for 10BaseT)
Ethernet Hub or Switch
Server or peer

8. Your LAN connected computer
Computer needs appropriate bus for NIC
for example, if it’s a PC:
ISA (Industry Standard Architecture) is passe
PCI (Intel’s Peripheral Component Interconnect) is modern
PC-Card (used to be called PCMCIA) for laptops.
Parallel printer port ( for old laptops).
Serial port (generally for external modem)
USB (Universal Serial Bus) is newest approach

9. The Network Interface Card (NIC)
NIC are the I/O device for NOS
Also known as LAN adapters
NICs Amplifies electronic signals
Packages data for transmission
Physically connects computer to transmission media (cable)
Associated with the NIC is a unique address called the MAC (Media Access Control) address
OSI Layers 1 and 2
PC or Workstation
loaded with OS and NOS
A circuit board which plugs into an expansion slot of a PC, workstation or server, whose function is to connect these to a Local Area Network. Designed to comply with both a specific LAN Medium Access Control procedure (e.g. CSMA/CD for Ethernet, token passing for Token Ring) and a specific physical medium (e.g. twisted pair wire, coax, or multimode fiber). Associated with the NIC is a unique address called the MAC address.
The Network Interface Card (NIC) provides network communication capabilities to and from a computer system.
In addition to network operating software, each network device must also have a network interface card. These cards today are also referred to as adapters, as in “Ethernet adapter card” or “Token Ring adapter card.”
The NIC amplifies electronic signals, which are generally very weak within the computer system itself. The NIC is also responsible for packaging data for transmission, and for controlling access to the network cable. When the data is packaged properly, and the timing is right, the NIC will push the data stream onto the cable.
The NIC also provides the physical connection between the computer and the transmission cable (also called “media”). This connection is made through the connector port. Examples of transmission media are Ethernet, Token Ring, and FDDI.
The next step in building the LAN is adding a wiring hub.
Network Interface
Card (NIC)
Network Connector Port

10. The Network Interface Card
Today, NIC generally means Ethernet NIC
Could be 10BaseT, 100BaseT if more modern, 10 or 100BaseF if fiber, or 1000BaseT - GigE
may do speed/duplex autonegotiation
10/100 Mbps, Full/Half Duplex
may offload some upper layer protocol processing from CPU
checksumming and CRC (Cyclic Redundancy Check) calculation
automatic decision to accept (in promiscous or sniffing mode) or reject (normally) packets not addressed for this workstation
multicast address recognition

11. The Wire Today, generally Category 5 UTP (Unshielded Twisted Pair)
4 pairs of copper wire, twisted and bundled to tight specification
properly installed, provides Mhz signalling bandwidth.
Ethernet and Fast (100 Mbps) Ethernet use two of four pairs.
Cat 5 Ethernet uses “RJ-45” connectors
In older installations, could also be Cat 3, 10Base2 (“thin-net”) or even 10Base5 (“thicknet”)
In newer installations: Cat 5e, Cat 6, or even Category 7.

12. Ethernet hub or switch Switch is a multi-port bridge Hub
provides dedicated bandwidth to each port
may provide VLANS, multicast services
Hub
fine for small LANs (under about 6 people)
shared bandwidth
Hubs normally require no configuration
Sophisticated switches need administration

13. Hub
Hub: Device that serves as the center of a star-topology network, sometimes referred to as a multiport electrical repeater, or in Ethernet, a wire concentrator; not intelligent
??: Who can tell me what a repeater does? How is a hub different?

14. Hubs: A Single Bus Hub 126 123 127 124 125 128 1 single Segment
Hubs operate very much the same way as a repeater. Amplifies and propagates signals received out all ports, with the exception of the port from which the data arrived.
For example, if system 125 wanted to send a message to system 127, the message would be sent to all systems on Segment 1, as well as across the bridge to all systems on Segment 2. System 127 would see that the message is intended for it and would take it. The printer would even receive the message, but would do nothing with it.
I should mention here that devices on the network are constantly listening for data and inflation. When it senses a piece of information that is addressed (and we will talk more about addressing later) for it, then it will accept that information into memory found on the network interface card (NIC) and begin processing the data.
1 single Segment

15. Hub Summary OSI physical Layer Amplifies signals
Propagates signals through the network
Must observe 4 repeater diameter limit
No filtering
No path determination or switching
Used as network concentration point
Single physical Collision Domain

16. Bridge: The Switch’s Predecessor
Very simple device that connects and passes packets between two or three network segments
Bridge is intelligent -- analyzes incoming packets and forwards or drops based on MAC Layer 2 addressing information.
Two kinds of bridging methods primarily--
Transparent Bridging (Ethernet)
Source-route Bridging (used in token-ring environments)
Bridges improve network throughput and operate at a more intelligent level than do hubs. A bridge is considered to be a store and forward device that uses unique hardware addresses to filter traffic that would otherwise travel from one segment to another. A bridge performs the following functions:
Reads data frame headers and records source address/port (segment) pairs
Reads the destination address of incoming frames and uses recorded addresses to determine the appropriate outbound port for the frame.
Uses memory buffers to store frames during periods of heavy transmission, and forwards them when the medium is ready.

17. Bridge Example BRIDGE 126 123 127 124 125 128 Segment 2 Segment 1
The bridge divides this Ethernet LAN into two segments, each connecting to a bridge port. Stations are on segment 1 and stations are on segment 2.
When station 124 transmits to station 125, the bridge will not forward the frame because it recognizes that stations 124 and 125 are on the same segment. Only traffic between segments passes through the bridge. In this example, a data frame from station 123, 124, or 125 to any station on segment 2 would be forwarded, and so would a message from any station on segment 2 to stations on segment 1.
When one station transmits, all other stations must wait until the line is silent again before transmitting. The most common network protocol used today is Ethernet. In ethernet, only one station can transmit at a time, or you will have what is referred to as collisions of data. We will go through ethernet in much greater detail later.
Bridges will listen to the network and keep track of who they are hearing. For instance, the bridge in this example will know that system 127 is on Segment 2, and that 125 is on segment 1. The bridge may even have a port (perhaps out to the internet) where it will send all packets that it cannot identify a destination for.
125
128
Segment 2
Segment 1

18. Learning • Source MAC addresses are associated with Ethernet ports
E0: c
E0: c
E1: c
E1: c
0260.8c
0260.8c
Port Eth. 0
Port Eth.1
0260.8c
0260.8c •
Source MAC addresses are
associated with Ethernet ports

19. Bridge Learning
The source MAC address of all frames received is compared against the information in the forwarding database.
If the source MAC address is not found in the forwarding database, it is added along with the port identifier it received on.
The age value of this database entry is reset to indicate that this is a "fresh" entry
If the MAC address was last seen on a different bridge port, then the port identifier for the entry is changed to the new port identifier, and the age value of this entry is reset.
If the frame is received on the same port as the port in the database entry, then only the age value of this entry is reset.

20. Forwarding • Traffic propagated to specific destinations Port Eth. 0

21. Filtering, Dropping X X • Congestion control through network knowledge
E0: c
E0: c
E0: c
E1: c
E1: c
0260.8c
0260.8c
Port Eth. 0
Port Eth.1 X X
0260.8c
0260.8c •
Congestion control through
network knowledge

22. Spanning Tree Removes Loops
Problem: a frame gets flooded or forwarded in both directions around a loop
Solution: makes one of the ports inactive until needed for redundancy
1) elect root
2) cheapest “port to root” for every bridge
3) cheapest bridge (designated) for every segment
4) block more expensive,unused port
Segment 1 B
Root
Segment 2 9 10 B B 10 B 10 10 10
Segment 3
Cost 10
Loop: the erroneous repeated traversal of the same path by the same frame of information. B 10
Segment 4
- Denotes blocked port

23. Switch? Hub? What’s the difference?
Hub Gives
Shared Bandwidth
Switch Gives
Dedicated Bandwidth

24. Switches Use bridging technology to forward traffic between ports;
Provides full dedicated data transmission rate between two stations that are directly connected to the switch ports
Builds and maintains address tables called content addressable memory (CAMs)
CAMs match MAC addresses to make rapid forwarding decisions

25. Switches are better than Bridges
10 Mbps
Coaxial Cable
“shared”
Workstation 34 31
BRIDGE 35 32
Now, let’s add some complexity to the network. Let’s say that you have someone who does a lot of video editing, on-line, and also stores the edits and clips in digital format on the server. Well, the server is on segment 1. Let’s also say that the capacity of the cable is 10Mbps.
In single-segment, repeated, and bridged environments, all stations on this Ethernet network share the 10-Mbps bandwidth. Superservers, high capacity workstations, and multimedia applications have fueled the need for higher capacity bandwidths.
In the example above, only one workstation is able to transmit data at a time. Consider the amount of bandwidth a video segment would consume if the station in the upper-right were to send a large file to the filer server for storage. Or, consider how much time would be consumed if one of the users were to send a very large print job to the printer.
Each of these activities consumes some of the 10-Mbps bandwidth in the LAN. 33 36
Segment 1
Segment 2

26. Switches are better than Bridges
10 Mbps
UTP Cable
“dedicated”
Workstation 34 31 35 32
SWITCH
Now, let’s replace the bridge with something similar, yet different in some key ways to that of the bridge, a switch.
An Ethernet Switch will treat every port as a separate Ethernet segment and use bridging technology to forward traffic between ports. In this manner, switching provides full 10-Mbps data transmission between two stations that are directly connected to the switch ports.
Let’s say that station 34 is in the process of saving a large file to the sever 33. This will consume 10-Mb of bandwidth. At the same time, lets say that station 32 wants to print. Both of these activities can happen simultaneously. Station 32 will have an direct connection with 36, the printer, while 34 is saving files to the server 33. Neither of the stations involved will experience any bandwidth problems, or have to wait for each other to finish their work before proceeding. Rather than the devices having to share the 10Mbps bandwidth, they will each have a “dedicated” line.
Now, there are still times when you may want to use a simple bridge, combined with switching functionality. Let’s say for example sake, that we have the flexibility to reorganize the LAN to incorporate our changes. 33 36
N Segments
N Segments

27. Switch Advantages Packets switched over dedicated links
Users experience better performance
LAN switches permits users to transmit simultaneously.
Assigning users bandwidth based on need.
Some users use 10 megabit ports and NICs
Some users and servers get 100 megabit ports
Full Duplex transmission possible

28. The Big Picture Dedicated 100 Mbps Shared 10Mbps Ethernet Hub Ethernet
Workstation
Dedicated 100 Mbps
Shared 10Mbps
Multimedia
Workstation
Ethernet
Hub
Ethernet
Switch
Server
Dedicated 100 Mbps
Router
Printer
Internet or
Corporate Intranet

29. Operating Systems and Network Operating Systems

30. The Operating System ties together all the computer’s hardwareHW HW HW HW HW HW HW HW
The physical equipment and components of which a device is constructed. Contrast with the instructions and algorithms to operate or control such devices (software). OS HW HW HW HW

31. OS Definition
The central control program on a computer that manages all aspects of the computer's hardware and controls the execution of software operations
Collection of programs which provide the computer with capabilities and functionality not specifically supported in the computer's hardware.
Usually linked closely to a vendor-specific computer architecture, called a platform. Examples include DOS, OS/2, Apple Mac OS, UNIX and Microsoft Windows 95/98 and Windows NT.

32. The OS coordinates all this
Fax/ Modem
Mouse,
Keyboard
Driver OS
NIC
Monitor
Driver is the software entity that provides an interface between the general software in a computer and device-specific hardware, such as:
Scanner

33. OS Features An OS: Examples of devices:
Makes devices available to applications via software
Allows users to interact with applications by using these devices
Examples of devices:
Input: mouse, keyboard, tablets, scanners
Output: printer, plotter, fax modem, phone line, monitor, HD, Backup Tapes, video camera.

34. Network Operating System (NOS)
If OS is the software that manages the resources of the computer, then
NOS is the software that manages the resources of the network
NOS models:
Peer-to-peer model
Server Based model
PCs, hub/Switch, printer servers, routers

35. Server Based NOS MS Windows NT (TCP/IP mainly)
Peer
Peer
Peer
MS Windows NT (TCP/IP mainly)
Novell’s Netware (IPX and TCP/IP)
Banyan Vines
IBM’s LAN Server

36. NOS Services File Service Messaging Service Data Base Service
Print Service
Messaging Service
Directory Service
Communication Service
Usually, NOS services are provided by one or more centralized servers
Services often administered by the IS department
File Service: makes remote files appear to be local
Data base service: Supports multi-user access to a centralized or distributed information store
Print Service: makes remote printers appear to be locally attached. Multiplexes print jobs from one or more users to one or more printers
Messaging service: provides a store-and-forward messaging system to route
Directory service: Maps mnemonics symbols (i.e., names) to specific resources and network addresses. Most directory services can also provide user authentication (I.e., NT user profiles) and support network resource allocation
Communication service: Supports dial-in, dial-out networks access. It also supports modem sharing making remote modems appear to be local

37. O/S and NOS
Modern Operating Systems have built in network drivers (OSI compliant stacks)
Network Operating Systems (NOS) support some form of network-oriented client/server application:
Novell Netware for PC’s (Client/Server)
Appletalk for Apple Macintoshes
NFS for UNIX systems
TCP/IP, NetBEUI, DLC, etc.. Systems under MS-OS

38. OS and NOS Examples - UNIX
First OS with integrated networking, based on IP
Unix services were inspiration for most other NOS vendors
NFS (Network File System) for file serving
LPR (line printer) for printer sharing
TCP/IP and UDP, etc. standard on all UNIX platforms
Examples of UNIX OS:
LINUX, HP-UX, IBM AIX, DEC Ultrix, SunOS, Solaris, SCO UNIX, BSD4.3, System5.4
Supported NICs: Ethernet, Token-Ring, Fast Ethernet, Gigabit-Ethernet, ATM
A flexible and powerful OS, designed to operate independently of any specific computing architecture. Deployed on a wide variety of computers worldwide, from desktop workstations through mainframes and telco switches. Developed and initially released in 1969 by Ken Thompson of Bell Laboratories.

39. O/S and NOS Examples : MS-DOS/Windows 95/98
Most popular O/S for 32-bit Intel compatible PC’s
Market share over 90%
Initially without network capabilities
Networking only possible with additional software: Novell Netware, Microsoft LAN Manager
Enhanced version Windows95/98, integrate networking capabilities
NetBEUI, TCP/IP, Novell IPX ODI Protocol, IBM DLC, Microsoft DLC, Banyan VINES protocol, ATM Call Manager, etc..
A multitasking operating system produced by Microsoft Corporation for 32-bit Intel compatible computer systems. The Windows 95 Program Manager uses either the default Explorer or the older Windows 3.x interface. It fully supports the Plug and Play hardware specification.
Designed as the successor to Windows 3.x, it became one of the fastest selling software packages in history within days of its formal rollout in 1995.

40. O/S and NOS Examples : Windows NT (Now improved into Windows 2000)
A 32-bit preemptive multitasking, multithreaded, multiprocessing operating system
Developed by Microsoft with the same w95/98 GUI
Initially deployed on Intel PC architecture, then ported to Motorola PowerPC, Digital Alpha AXP, and MIPS R4000 platforms
Brings Windows closer to UNIX performance
Two level approach: Windows NT Server 4.0 and Windows NT Workstation
Extended HTML/Java Services
NOS: mainly TCP/IP based
All Kind of NIC cards and protocols supported
A 32-bit preemptive multitasking, multithreaded, multiprocessing operating system developed by Microsoft Corporation. Initially deployed on Intel PC architecture, it has since been ported to Motorola PowerPC, Digital Alpha AXP, and MIPS R4000 platforms.
It uses either the standard Windows 3.x or Windows 95 Explorer GUI. It is available in Workstation and Server versions, and has been increasing in popularity rapidly since its introduction in 1993.

41. LAN software and client configuration
Windows Networking, pre 2000, minimizes configuration, but can’t be routed globally.
Workstations learn about neighbors by broadcasting identity
TCP/IP networking must be explicitly configured
Allows global connectivity from PC Host
Dynamic Host Configuration Protocol (DHCP) solves configuration problem
Windows 2000 seeks best of both worlds

42. Windows Networking (before Windows 2000)
Workstation can be locally administered
Don’t need (or get) globally recognized name or number (like an IP address).
NetBIOS (used by Windows) is not a routable protocol
Windows broadcasts your computer’s name so others can locate it.
If necessary, locates a Domain Controller to authenticate username and password
Uses WINS server (Windows Internet Name Service) to resolve other workstation names

43. TCP/IP host configuration
Need to configure computer with:
IP address:
Subnet size or “mask”:
Default router IP address:
use this gateway to access hosts outside this subnet
DNS (Domain Name Server) IP address
can be located outside local subnet
Domain name: lucent.com or sales.lucent.com
for locating local machines, domain name is added
for example, vail.lucent.com could be called “vail”

44. TCP/IP host configuration - DHCP
Configuration is complicated, so DHCP (Dynamic Host Configuration Protocol) was invented
Client broadcasts DHCP requests
DHCP server responds with:
Unique IP address for client to use
correct subnet mask
default router, DNS server, Domain name
Other services: WINS very common
DHCP provides client with everything needed.

45. Windows 2000 Networking Uses TCP/IP, but with a twist
Tries to maintain Windows “ease of use”
DHCP to configure workstations
Dynamic DNS to name workstations
Active Directory powerful way to replicate data between servers.

46. The path to Client/Server
Mainframe Architecture
File Sharing Architecture
Client Server Architecture

47. Mainframe architecture
Not a client/server architecture.
All intelligence is within the central host computer
Users interact with the host through a terminal
Terminals send keystrokes directly to the host
User interaction can be done using PCs and UNIX workstations
Older mainframe software architectures do not easily support graphical user interfaces (GUI)

48. File sharing architecture
Not a client/server architecture
The original PC networks were based on file sharing architectures
The server downloads files from the shared location to the desktop environment
The requested user job is then run (including logic and data) in the desktop environment
Limitations: works if shared usage is low, update contention is low, and the volume of data to be transferred is low.
In the 1990s, PC LAN computing changed because the capacity of the file sharing was strained as the number of online users grew and graphical user interfaces (GUIs) became popular
PCs are now being used in client/server architectures

49. Client/Server architecture
As a result of the limitations of file sharing architectures, the client/server architecture emerged
Database server replaces the file server
Using a relational database management system (DBMS), user queries could be answered directly
Reduces network traffic by providing a query response rather than total file transfer
Improves multi-user updating through a GUI front end to a shared database
Remote Procedure Calls (RPCs) or Structured Query Language (SQL) statements are typically used to communicate between the client and server

50. Client/Server The future
Two tier architectures
good solution 10 to 100 people interacting on a LAN simultaneously by maintaining a connection via "keep-alive" messages with each client, even when no work is being done
Three tier architectures (multi-tier architecture)
middle tier was added between the client and the database management server environment. Good for large number of users (in the thousands)
Distributed/Collaborative Enterprise Architectures
emerged in 1993 Based in Object Request Broker (ORB) and Common Object Request Broker Architecture (CORBA)
Based on Object Oriented data bases (C++, SmallTalk, JAVA)
Two tier architectures
good solution 10 to 100 people interacting on a LAN simultaneously
This limitation is a result of the server maintaining a connection via "keep-alive" messages with each client, even when no work is being done
Three tier architectures (multi-tier architecture)
middle tier was added between the user system interface client environment and the database management server environment
large number of users (in the thousands)

51. The Client/Server Basics

52. The Client
Application sitting in a PC or workstation, along with its associated software, actually operated by an end user.
Some important attributes of the client include:
the applications software installed on the computer;
the client's user interface, which provides a method for interacting with the server (back office).
the computer's processor, memory, and networking options.

53. The Server
Back end Application (called daemon in Unix) installed in a central PC, workstation, minicomputer or large computing system
Provides a central service to user stations on a Local Area Network or across the WAN
Typical examples of these centralized services are:
Data bases
File storage (file servers),
Printer sharing (print servers),
Centralized application software storage/processing (application servers),
and Wide Area Network access (communication servers).

54. Compact Definition
Client server is referring to a model which divides applications in two parts:
Client or front-end: the software process used by the end-user
Server or back-end: the software process created by the actual application running in the server
On a machine you can find multiple clients and multiple servers at the same time
One of the first problems you will encounter talking about clients and servers is the fact that people are not always very clear on what they talk about.
In the strict sense a client is nothing than a software process running on a computer. It is a client because it uses resources from somewhere else. The somewhere else is the server side. A server therefore again is nothing than a software process.
In quite a few instances people will talk about a client as being a piece of equipment e.g. a PC or a workstation. Big machines are often referred as servers.

55. An Example: Checking Bank account
Authentication process for security
The client program in your computer forwards your request to an intermediate server program at the bank
That program may in turn forward the request to its own client program that sends a request to a database server at another bank computer to retrieve your account balance.
The balance is returned back to the bank data client, which in turn serves it back to the client in your personal computer, which displays the information for you.

56. Client/Server and Internet
Web browser is a client program that requests services (the sending of Web pages or files) from a Web server (which technically is called a Hypertext Transport Protocol or HTTP server) in another computer somewhere on the Internet.
Your PC with TCP/IP installed allows you to make client requests for files from File Transfer Protocol (FTP) servers in other computers on the Internet.

57. Drivers To adopt Client/Server
TCO Model - Total Cost of Ownership is lower
Central high performance is cheaper than distributed performance.
One fast printer is cheaper than 10 slow
One large fast file server is cheaper than 100 medium sized disks in PC’s
One fast computer server is cheaper than medium-fast PC’s or workstations (TCO)
The key driver for client-server is money. It is more efficient, easier to manage and in most cases also cheaper to centralise resources. One fast shared laser printer on the network is a lot cheaper than a laser printer connected to each individual PC.
Oracle
Financial
SAP
PeopleSoft

58. Server Based Model All communications centers around the server
Client-to-server
Server-to-server
Client is a user of services, Server is a provider
Clients do not directly communicate with other clients
85% of the marketplace
Servers are a high-powered micro, mini or mainframe computer

59. Server Based Pros and Cons
Scalable
Ease of management and security
Single point of administration and maintenance
Easy to to provide fault-tolerant services
Redundant power supply, mirrored servers, back-up
Advanced services
Cons:
Difficult to install
Requires staff expertise

60. Peer-to-Peer Model
Allows communication between any two devices in the LAN
Good for small networks environment
Pros:
Flexible
Each workstation can provide services to other devices
Easy to install and run
Basic services: file sharing, printer sharing, services
Cons:
Difficult to administer when network begins to grow
Do not offer a central point of administration

61. Peer-to-Peer Apple Talk Windows 95/98 using NetBEUI
Novell Personal Netware
Napster!

62. Napster is a big Peer-to-Peer network
Other Internet-wide file sharing services are coming too
Freenet, Gnutella, FileFury
All of Internet becomes as easy to access as the LAN.
Directory Servers automate administration
Capitalizes and insists on high speed Internet and WAN connections.
Servers on WAN seem as “local” as the LAN.

63. The Future of Local Area Networks
Today, LAN is gateway to Internet and WAN
Wireless, IP telephony latest in a series of important LAN drivers
Ethernet (and soon wireless) only way to go
Fast Ethernet and Gigabit Ethernet are easy upgrade paths for older networks.
Client/Server will continue to evolve
Agents, Java, B2B, B2C.
Web-based computing paradigm keeps older networks, computers viable while driving new application and software innovation.

64. Food for Thought
What are the advantages or disadvantages of a Web server/client architecture?
What’s the difference between the mainframe/terminal relationship, the client/server relationship, and the web server/web browser relationship?