Tuesday, December 30, 2008

MULTIPLEXING

Multiplexing the basis of a data allow many relationships network.Multiplexing (connections) over the network share the same transmission facilities. Two main types of multiplexing are discussed here; Time-division multiplexing (TDM) and Statistical multiplexing (statmux).

TDM TDM

Time-division multiplexing is allocated a certain amount of time in a physical circuit to a number of connections.. Because of the physical circuits typically have data flow velocity is constant, then the allocation of the amount of time in the circuit is equal to the bandwidth allocation.

TDM is the synchronization technology. Data is transmitted into the network to the primary time source (master clock), so that there would be no traffic jams when the data is transmitted.

One of the main problems of TDM is the bandwidth allocated to a number of connections is only allocated to those connections, whether they're used or notSo we are still paying for unused capacity, this result is quite expensive TDM.

Statistical Multiplexing Statistical Multiplexing

Statistical multiplexing become popular due to cost issues in TDM. Statistical multiplexing transmission bandwidth is shared between all users of a network, without a dedicated one reserved for the connection.

One of the benefits of TDM statmux is cheaper. With statmux network, we can sell more capacity than the network we have. In theory, not all users want the data transmission network with a maximum speed at the same time.

There are several technologies statmux, there are three main things of interest in this technology in the past 10 years are; IP, Frame Relay, and ATM. The MPLS can be said is the fourth type of statmux technology.

Statmux technology works by dividing the network traffic into discrete units and handle all the units separately. In IP, the unit is called a packet; on Frame Relay is called frame; the ATM cell is called. The above items have the same concept in each case. Stamux networks allows carriers handle more than that owned by the network used (oversubscription), are much cheaper than the TDM circuit.

Issues - issues that are in Statmux

Statmux introduce some things that are not found in TDM networks. When the packet enters the network is not simultaneous (asynchronous), it resulted in conflicting sources. If two packages right into the router at the same time (coming from two different interfaces) and addressed to the same output interface, it is a source of conflict. One of the packages have to wait another packet to be transmitted, while packets are not transmitted to wait until the first packet sent on the link in question. Although the time delay occurs, but usually not large.

There are also some things to do with the package contained in the buffers. Some types of traffic (bulk data transfer) are described by way of stored (buffered) while the other traffic (voice, pictures) do not. Necessary so that different treatment mechanisms to meet the demand for different applications used in network.

Statmux technology has 3 abilities that are not owned by TDM technology, namely; buffering (storage), Queuing (queue), and Dropping (Decrease).

Frame Relay has a simple method that can answer these issues. The concept is a committed information rate (CIR) , forward and backward explicit congestion notification ( FECN and BECN ) and discard

IP has a Diff Serv Code Point (DSCP) bits, which evolved from the primary IP bits. IP also have the random early discard (RED), which has the advantage that the TCP both in terms of handling the drop and TCP is the transport-layer protocols of the most widely used for IP.. Finally, IP has a bit of explicit congestion notification (ECN), which is still fairly new and only used a limited basis.

ATM explain the source of the conflict by dividing the data in a small size called cells. ATMs also have the 5 different classes of service, namely:

  • CBR (constant bit rate) CBR (constant bit rate)
  • rt-VBR (real-time variable bit rate) rt-VBR (real-time variable bit rate)
  • nrt-VBR (non-real-time variable bit rate) nrt-VBR (non-real-time variable bit rate)
  • ABR (available bit rate) Abr (available bit rate)
  • UBR (unspecified bit rate) UBR (unspecified bit rate)

IP is a protocol statmux first. RFC 791 defines IP in 1981 and became important in recent years. Frame Relay is not available commercially until the early 1990s, while the ATM appeared in mid-1990.To view or download this please RFC 791 to: http://www.ietf.org/rfc/rfc0791.txt?number=791.

One of the problems faced by network administrators when replacing TDM circuits to Frame Relay and ATM circuits is that running IP over FR or ATM is a protocol running over statmux other statmux protocol. This is usually less than optimal because the existing mechanism in the layer statmux to resolve conflicting resource issues are often not translated perfectly to the other.

That's what the considerations for choosing one of two things. One of them is to avoid congestion at the network layer of the two statmux, or we find a way to map the three layers of conflict regulation mechanism to the second layer of conflict regulation mechanism. Because those two things is something that is not possible and it is not financially attractive to avoid congestion at the network layer statmux two, we need to be able to map the three layers of conflict regulation mechanism to layer two. This is one reason MPLS plays a key role in the development of current networks.

What is Traffic Engineering?

Network engineering be manipulated to suit the network We made the best predictions of how the traffic can go (flow) through the network so that we can choose the appropriate circuits and network equipment (routers, switches, etc.) as appropriate. Engineering Network is usually done in the long run because the time required to install the circuit or new equipment can be very long. Traffic engineering to manipulate traffic in accordance with the network. No matter how persistent traffic in the network tried would not fit 100% with a prediction that has been made. Traffic Engineering is essentially moving traffic so that traffic from the links that have the congestion moved to a link that is not in use. Traffic Engineering can be implemented in a way that is as easy as tweaking IP metrics in the interface or something as complex as running a full-ATM PVC mesh and PVC path optimizing based on past traffic demand.

Traffic engineering with MPLS is an attempt to obtain the best connection-oriented traffic engineering techniques (such as the placement of ATM PVC) and combined with IP routing (IP routing). The theory is to do t raffic engineering with MPLS is more effective as the ATM, but without the many shortcomings of IP over ATM.

MPLS traffic engineering before

IP traffic engineering in general control the point where the IP through our network to alter the usual relationship. There is no possible way to control the path of traffic based on where traffic is coming, but we can only control where the traffic destination. Although both IP traffic engineering and many large networks to use it successfully, we will see that a lot of IP traffic engineering problem that can not be resolved.

ATMs can be used to pass the PVC in the network traffic from a source to the destination. This means we have more things that controlled the flow of traffic in the network. Some of the largest ISPs in the world using an ATM to control traffic on the network. They do this by making an ATM PVC between a set of routers and periodically re-measure and placing it on the ATM PVC traffic study of the router. But the problem that arises is that the router is a full-mesh causes O (N2) will flood when a relationship (link) die and O (N3) when the router dies flooding.This causes much concern in several large networks.



No comments:

Post a Comment