|
FALL 2000 Lecture Notes 11: NETWORK MANAGEMENT A few beginning concepts – you probably know this stuff – but it won’t hurt to review. Networking is a means of connecting computers together so they can share- Data, such as files and programs Each computer or workstation on a network is called a network station, a node, or possibly a host. Each processor classifies other processors and their resources as remote, and considers its own resources as local. Typically, a host at one site, called the server, has resources that a host at another site, called the client, wants to use. Hosts can alternate being clients and servers. Data is sent over the network as bits and bytes Data to be sent over the network is first divided into segments, each with a header and trailer - A header is information sent in front of data to identify the data destination and protocols Data packets or frames are sent as independent units over the network. At the receiving end, the header and trailer information is removed, and the data is reassembled.For PCs, the network card breaks the data into segments and adds the header and trailer, creating the packets. LAN (local area network)
WAN (wide area network)
The 3 most popular physical network architectures, or topologies,are - Ethernet Architectures are distinguished basically by - How computers are logically connected Ethernet is the most popular network topology used today. - Bus network- connects each node in a line with no centralized point of connection. The star is the more popular Ethernet configuration. The Token Ring is logically a ring, but stations are physically connected to the network in a star formation. - Each station connects to a centralized hub called a controlled-access unit (CAU), multistation access unit (MAU), or smart multistation access unit (SMAU). FDDI (Fiber Distributed Data Interface) A ring-based network without a centralized hub, making it a physical and logical ring. Uses a token-passing method to control traffic, but is more powerful and sophisticated than Token Ring. Stations can pass more than one frame along the ring without waiting for the first to return. Possessing the token gives a station the right to transmit. Data frames travel along the ring without the token. A station holds the token until it has sent its data, then passes the token on. More than one station can have frames on the ring at the same time.Network Interface Card (NIC) - Plugs into a systemboard.
TRANSPORT PROTOCOL STANDARDS OSI Layer Network Model International Organisation of Standardization (makes recommendations about data communications) took on the task of creating such a network architecture (OSI reference model). Is a framework for defining services network should provide to users. The word "open" means that any two systems that conform to the reference model and the related standards can be connected, regardless of the platform. Once all services were identified, similar clusters were collected together in layers. Reason for doing this was to group easily localised functions so that each layer could be re-designed and its protocols changed in any way to take advantage of new advances without changing the services to and expected from other layers. Boundaries between layers were selected at points that past experience suggested were effective. Data is segmented into frames or packets for transmission. Each layer in the seven layer model can add information to the beginning and ending of a frame, as below: When the frame is presented to the receiving station, that layer interprets the information in the header and trailer intended for that layer, and then strips off the information before passing the frame onto the next higher layer in the model. For frames to transmit successfully, each layer of the OSI model must communicate with the same protocol as its counterpart in the remote computer. LAYER 1 The Physical Layer This is the protocol for how the data is transmitted.Sees the data as only bits that need to be passed on. This is the only layer concerned with hardware. The packaging of data has already occurred before this layer ; the physical layer does not distinguish the header from the rest of the package – just see all of it as bits that need to be passed on. LAYER 2 The Data Link Layer Responsible for receiving frames of data from the Network layer and splitting them into segments of bits to be presented to the Physical layer for transport. Software is needed to implement this layer. The data link layer establishes and controls the physical path of communication before sending data to the physical layer. Error checks in both directions. Looks for errors in transmission, and damages, lost or duplicate message frames. Token Ring or Ethernet firmware on network cards are examples of code that handles Data Link and Physical Layers.LAYER 3 The Network Layer Software at this level accepts blocks of data from Layer 4, the transport layer, resizes them to shorter packets and routes them to the proper destination.Responsible for finding the best route by which to send frames over a network Addressing methods that allow a node and its network to be identified, as well as algorithms to handle address resolution are handled. The most common protocols are TCP/IP (Transmission Control Protocol/Internet Protocol) and IPX/SPX (Internetwork Packet exchange/ sequenced packet exchange). For TCP/IP, the TCP portion is responsible for error checking, and therefore operates in the Transport layer and the IP portion makes up the network layer and is responsible for routing. , (we will talk more about these later). WinNT and Win98 support both protocols, Novell supports IPX/SPX. LAYER 4 The Transport Layer Maintains reliable data transmission between end users. This layer also has a mechanism to ensure that a fast host cannot overrun a slow terminal.LAYER 5 The Session Layer Responsible for establishing and maintaining a session between two networked stations or nodes.A session between two nodes in a network is sometimes called a socket. Transport layer and session layer are very similar, and OS of host computer handles session layer they are often combined into one as in TCP/IP. Layer 6 The Presentation Layer Receives requests for files from the Application layer and presents the requests to the Session layer.Responsible for data manipulation functions – formatting, compression and encryption. Also data conversion, syntax conversion and protocol conversion. Accommodates totally different interfaces as seen by a terminal in one node and expected by the application program at the host computer. Layer 7 The Application Layer Responsible for interfacing with the applications software that is using the network.Responsible for formatting user data before passing it on to the lower layers for transmission. It contains network management functions such as file transfer and e-mail. NETWORKING SOFTWARE Each layer in the OSI model uses a different method of communicating to its counterpart layer
The 3 best-known PC network operating systems are UNIX, NetWare by Novell, and Microsoft Windows NT Network protocol: Point-to-Point Protocol (PPP) is a protocol by which PCs with modems can connect to a networkIt is the most popular protocol for managing network transmission from one modem to another At the Network and Transport layers, TCP/IP is a protocol used by the Internet E-mail is sent using SMTP (Simple Mail Transfer Protocol), which operates on the Session layer. Network drives use NFS (Network File Service) protocol, and the World Wide Web uses HTTP (Hypertext Transfer Protocol), both on the Session layer. Transferring files across the Internet is usually done using FTP (File-Transfer Protocol) Network services provided at the application and presentation levels are:
MANAGEMENT OF NETWORK FUNCTIONS Early networks focussed on sharing expensive hardware such as large mainframes, laser printers and sizable hard disks. System owners soon realized that the real value of a network was not the hardware but the information. Operating systems were enhanced with network capabilities. Network operating systems include networking as an integral part of the operating system, not as an extension of it. A network operating system has to address the issues of memory, device, processor and file management for the network. MEMORY MANAGEMENT M.M. tracks available memory on each node by use of kernel that performs paging based on the goals of the local system. Policies and mechanisms implemented at the local sites will be driven by global system requirements. At the network level, memory manager receives requests from the process manager to provide memory to new or expanding client or server processes. MM’s function extended from that of controlling the local machine to accept requests for memory from both local and global sources. Combined memory of entire system made up of several subpools, one for each processor – network MM has subcomponent that exists on each machine. PROCESS MANAGEMENT The creation of a process by a network operating system involves the creation of a PCB with similar info but with additional info to indicate the location of the process in the network. The deletion or termination of a process requires finding the PCB and having the authority to access and delete it. High level of cooperation maintained in system sites in determining which process should be loaded and where it should be run. Once the process has been scheduled for service, must be initiated at assigned site, requiring a dispatcher, which allocates the device to the process and initiates its execution. This may even involve:
Interrupts may require service from IO outside the site and the process manager in a network system would have to handle this. DEVICE MANAGEMENT Performed at the physical level same as for workstation OS. Network OS must keep a global accounting of devices and their availability, maintaining each devices status and control blocks and distributing this info to all sites. Systems are configured around server processes, which manage multiple resources. A server process consists of device driver, device manger and network server component. FILE MANAGEMENT Goal of networked file management is to give illusion of a single logical file system implemented an assortment of devices and computers. Network OS must manage the files in the same way as single user OS. Should read, write, open, close, create and delete files. However, all potential users of files must be locked out of files being currently used by other users. NETWORK MANAGEMENT Interprocess communication is transparent to the users of the network. The Network manager has the function of controlling the allocation of ports to processes, identifying the processes in the network, controlling the flow of messages. Relieves users of having to know where processes physically reside in the network. Network manager accepts each process’s send and receive commands and interprets them, transforms these commands into low level actions that perform the transmission of messages over the links.
WinNT Networking Windows NT Server Has all the functionality of Windows NT Workstation - Offers the additional functionality of administering and monitoring the network from a centralized locationBoth Windows NT Workstation and Server can be configured to work as one node in a workgroup or as one node on a domain. Workgroup - A logical group of computers and users that share resources Domain A Windows NT domain is a group of networked computers thatShare a centralized directory database of user account information and security for the entire set of computersThe backup domain controller (BDC) is a computer that has a backup of the directory database The primary domain controller (PDC) stores and controls a database of - User accountsThis database is called the directory database or the security accounts manager (SAM) database. Used on PCs to control who has access to what programs, files, and other resources on a PC or network With DOS and Windows 95, there is no all-encompassing security to the PC except the power-on password Windows NT does provide all-encompassing security to the PC User accounts, in a workgroup, must be set up on each computer User accounts, in a domain, can be set up from the centralized domain server During Windows NT installation, the administrator account is always created An administrator has the rights and permissions to all computer software and hardware resources Permissions granted to a user and the user’s OS environment are controlled by the administrator, who can - Assign restrictions and rights to a group
The Internet - The largest network in the world Most people access the Internet through an Internet Service Provider (ISP) that gives you - A username and password Each node on a network is assigned a unique IP (Internet Protocol) address - The address is made up of four numbers separated by periodsThe largest IP address is 255.255.255.255 IP addresses are divided into three classes Class A license - Assigned a single number to be used in the first octet of the address, which becomes the network address Class B license - Assigned numbers for the first two octets Class C license - Assigned three octets as the network addressStatic IP address - IP address is permanently assigned to a workstation Dynamic IP address - The server assigns an IP address to a workstation when it comes on-line to be used for the current session only A few IP addresses are reserved for TCP/IP Because of an impending shortage of IP addresses, a new scheme of IP addresses is being developed - Called IP version 6 (IPv6) Domain names are word-based names that can be used to address a host on the Internet, rather than the IP address A host computer with a certain domain name, connected to one network and assigned a certain IP address, can be moved to another network and assigned a different IP address The domain name can stay with the host while it connects to either network Two name resolution services track the relationships between domain names and IP addresses - Domain Name System (DNS) TCP/IP - Designed to enable routing of data from one network to another over a huge geographical area and over a huge number of networksTCP/IP can use more than one protocol at each OSI layer it supports When data is being transmitted, only one protocol is used at each layer of the OSI model, depending on the type of data being transmitted and the software interfacing with the network
|