Define
common networking terms for LANs and WANs. |
Compare
a file-and-print server with an application server |
File Server |
Centralised storage for files needed by a user group.
Benefits:
 | Centralised location |
 | Power conditioning / UPS |
 | Consistent data archiving / backups |
 | Speed (over peer-to-peer) - server usually higher spec |
|
Print Server |
Receive print requests from all networked devices, put in
a queue and spool to relevant printer. Some printers can have there own Network Interface
Card and therefore directly connect to the LAN. Still controlled by a print server - just
not directly tied to its parallel port. |
Application Server |
Actually host application software. To run an application
the client establishes a connection across the network to the application which runs on
the server.
Note: Servers that allow clients to download application software to run on the client are
acting as file servers where the files are application software. |
|
Compare
user-level security with access permission assigned to a shared directory on a server |
Share Level Security |
(Password protected shares) Individual resources given
access passwords. Useful on small networks (i.e. Peer-To-Peer) but cumbersome on large
networks with more resources and more passwords to remember. |
User Level Security |
User has an account to logon to the network. Permissions
to access resources are assigned to users on an individual basis. User only needs to know
one password. Benefits from higher security, central administration. |
|
|
Compare
a client/server network with a peer-to-peer network |
Peer to Peer |
Unstructured access to network-attached resources. Each
device can be a client and server simultaneously. Every machine is a peer of every other.
Benefits:
 | Inexpensive to operate - lack expensive servers and admin staff |
 | Familiar operating systems: Windows 3.11 / 95 / NT |
 | Can be more fault tolerant - server can be c |
 | Considered single point of failure |
Limitations:
 | Users must maintain multiple passwords - typically one for each resource
they need to access |
 | Burden for administration on each user - need good admin policies:
backups, filenaming, storage locations etc |
 | Security distributed throughout network |
 | Technical proficiency varies amongst users - security only as valid as
weakest peer |
 | Performance - each machine both client and server |
Recommend:
 | Max 10 computers in peer-to-peer network |
|
Client / Server |
Frequently shared resources consolidated onto a separate
tier of servers. Typically do not have a primary user. Benefits:
 | Security - server can be made more secure (locked away) and use of user
level security (centrally managed, only one logon password) |
 | Performance - each client works for its primary user only; server can be
configured for optimised performance (faster processor, disks, more memory etc) |
 | Administration - centrally co-ordinated backups, storage locations etc
|
 | Scalability - size of network limited only by hardware |
Limitations:
 | Costs more to implement and operate than peer-to-peer |
 | Loss of a server can impact all users of the network |
|
|
|
Compare
the implications of using connection-oriented communications with connectionless
communications |
Connection oriented |
Establishes a path between the two computers before
sending. Assures reliable delivery, i.e. phone conversation |
Connectionless |
Just sends packets, no checking whether they get there.
Fast but no assurance of delivery, i.e. letter mail |
|
Distinguish
whether SLIP or PPP is used as the communications protocol for various situations |
SLIP (Serial Line Interface
Protocol) |
Developed to provide dial-up IP connections in Unix
 | TCP/IP only |
 | Does not allow for dynamic IP addressing |
 | No software compression |
 | No encryption or password authentication |
|
PPP (Point to Point Protocol) |
Default option in Windows NT Dial-Up Networking. Newest
packet framing protocol that provides self-configuring, full-duplex, bi-directional,
peer-to-peer connections.
 | Allows for dynamic IP addressing |
 | Can negotiate header compression |
 | Supports error control |
 | Offers encryption and password authentication |
 | Supports NetBEUI, TCP/IP, IPX/SPX |
|
|
Define
the communication devices that communicate at each level of the OSI model |
Layer |
Description |
Application
7 |
This layer is used for those applications that are specifically written
to run over the network. Example applications such as file transfer, terminal emulation,
electronic mail, and NetBIOS-based applications. Protocols: DNS,
SNMP, TELNET, FTP, SMTP, NFS, SMB
Devices: Gateway (can work at all layers) |
Presentation
6 |
Data formatting, character code conversion and data encryption. This layer is responsible for data translation (format of the data) and data
encryption (scrambling and descrambling the data as it is transmitted and received). It is
not always implemented in a network protocol.
Redirector sits here. |
Session
5 |
Negotiation and establishment of a connection with another node.
This layer establishes, maintains and disconnects a communications link
between two stations on a network. Determines if mono- or bi-directional and ensures one
request is completed before a new one is accepted. |
Transport
4 |
Provides reliable end-to-end delivery of data beyond local LAN segment.
Can detect packets discarded by routers and automatically generate a
retransmit request. Re-sequences packets that arrive out of order before passing them to
the Session Layer.
Protocols: TCP, UDP, SPX, NetBEUI. |
Network
3 |
Routing of packets of information across multiple networks. Lacks any native error correction/detection and so relies on Data Link Layer for
end-t-end reliable transmission. This layer used to establish comms with computer systems
beyond the local LAN segment. Has own routing addressing architecture (different from Data
Link Layer machine addressing).
Protocols: IP, IPX, AppleTalk, ARP, RARP, RIP, OSPF
Devices: Router |
Data link
2 |
Responsible for providing end-to-end validity of data being
transmitted. Packs data into frames. Has own address architecture, applicable only to
other networked devices that reside locally. Successful delivery achieved by originating
node receiving an 'ACK' sent by destination node to indicate frame received correctly.
Errors detected and corrected via CRC check. IEEE split this
layer into Media Access Control (MAC) and Logical Link Control (LLC) layers.
Protocols: SLIP, PPP, DLC
Devices: Bridge, Switch |
Physical
1 |
Accepts frames of data from Layer 2 and transmits their structure
serially, one bit at a time. Does reverse on receive. Defines the
methods used to transmit and receive data on the network. It consists of the wiring, the
devices that are used to connect a station's network interface controller to the wiring,
the signalling involved to transmit/receive data.
Protocols: IEEE 802.2
Devices: Repeater |
|
|
Describe
the characteristics and purpose of the media used in IEEE 802.3 and IEEE 802.5 standards. |
Common features |
Common basis for hardware level addressing, management and
monitoring - permits mixed topology networks. |
802.3 (Ethernet) |
Contention based media access - devices compete for the
right to transmit. Dominant form now CSMA/CD (Carrier Sense Multiple Access with Collision
Detection). If a device detects a collision, it backs off, waits a pseudo-random period
and then tries again. Only one device can transmit at a time, half duplex. Logical bus network that can transmit data at 10Mbps. Data is transmitted on the
wire to every computer. Only those meant to receive the data acknowledge the transmission.
 | All devices on a segment share a collision domain (share same bandwidth).
Can be improved by a switch which gives a collision domain of two - the port and the
device connected to it. Device has full bandwidth to itself. |
 | Probabilistic (cannot state when a packet will reach its destination - a
collision may occur) |
 | Baseband transmission (digital signalling over a single frequency)
|
 | Media: thicknet (10Base5), thinnet (10Base2), UTP (10BaseT) |
Fast Ethernet (100Mbps) 802.3u |
802.5 (Token Ring) |
Token based media access - regulates access by passing
tokens, only the device holding the token can transmit. Logical
ring (actual ring is in hub) that transmits at either 4Mbps or 16Mbps.
 | Deterministic (no collisions, so can accurately determine max delay that
can be experienced) ideal for applications that require predictable delays. |
 | Baseband transmission |
 | Shielded and unshielded twisted-pair cabling |
 | Uses hubs known as MAU, MSAU, SMAU |
 | A 16Mbps network will not accept a 4Mbps network adapter card |
|
802.12 (VG-AnyLAN) |
Demand priority media access - central repeater 'polls'
ports connected to it. Not really in use - failed in the market place ! |
|
Explain
the purpose of NDIS and Novell ODI network standards |
NDIS (Network Driver Interface
Specification) |
Developed by Microsoft and 3Com - defines interface
between the network transport protocol and the Data Link layer network adapter driver.
 | Permits any NDIS-compliant protocol stack to operate with any
NDIS-compliant adapter driver. |
 | Defines a method for binding multiple protocols to a single driver or one
protocol to more than one adapter. |
|
ODI (Open Data-Link Interface) |
Serves the same function as NDIS for Novell NetWare and
Apple environments. Provides support for multiple protocols on a single network adapter
card. |