R.S.SINDHUTHEJA* et al.

[IJESAT] INTERNATIONAL JOURNAL OF ENGINEERING SCIENCE & ADVANCED TECHNOLOGY

ISSN: 22503676
Volume - 2, Special Issue - 1, 37 40

A MODIFIED ENERGY CONSERVING SWITCH ARCHITECHURE FOR LANS

R.S.Sindhutheja1, T.V.Vamsikrishna2, Ch.Ravisankar3, S.Jhansi4
Asst. Professor, ECM Department, K L University, Guntur Dt., A.P, India. Graduate Student, ECM Department, K L University, Guntur Dt., A.P, India, vamsiecm@gmail.com 3 Graduate Student, ECM Department, K L University, Guntur Dt., A.P, Indi,. ravisankar2008still@gmail.com. 4 Graduate Student, ECM Department, K L University, Guntur Dt., A.P, India.
2 1

Abstract
In todays world energy conservation has become very important ever thing is going under some alternative methods in order to save the energy like architectures and some eco-friendly techniques .now networks is very important in todays life every data is going through it only especially in Lans energy conservation for switch is very important. Previous work on reducing energy consumption has primarily looked at either changing link rates or putting interfaces to sleep. Due to the unpredictable nature of traffic, the energy savings achieved have been modest, do not scale, and incur losses and delay. In this paper we proposed many techniques so that we can conserve energy like Sleep, Awake, and Hibernate algorithms. The main extension of this paper is Hibernate so that the interface awake for long time can go to hibernate state and with zero packet loss. We obtain over 92-93% energy saving in the LANs networks.

Index Terms: Hibernate, Sleep and Awake. -------------------------------------------------------------------- *** ------------------------------------------------------------------------1. INTRODUCTION
The research activity on the Conserving Energy on the LANs from the past few decades is increasing. Technological reasons for reducing energy Consumption includes the increasing cost of cooling and the limits placed on increasing the speeds of routers and switches due to heat dissipation in the hardware. Suppose if a data Packet is lost in the internet network we have to get it again this take a lot of energy to get data packet again so we have to overcome this problem. In the previous papers they have used much architecture like merging networks to overcome the data packet problem. And some of the algorithms like sleep and awake helps the interfaces to sleep so that we can conserve energy in the networks. This paper develops a new solution i.e. is Hibernate state to the interfaces. When the network is from the packets it will go to the Sleep state and the packet is ready to come it goes to the awake state but it is not bringing total conservation. So that we introduce a new state called Hibernate to overcome he problem. Hibernate sate is one in which we can conserve energy more than in Sleep state the process will be explained in below sections. In the coming sections we give a literature survey on the merge networks architectures in section II. A survey on the Sleep and Awake algorithms in the sections III. A newly implemented Hibernate state which is better than the Sleep is given in the section IV. We gave further development in the section V.

2. SECTION II
In section we use to give a detail explanation of Merge networks. Which is an important architecture in order to conserve energy? When the data packets are coming from the internet typical switch will use to provide the interfaces to reach the destination. Suppose if any data packet is loss again its take a lot of energy to get the data packets from it. So merge network is one of the important architecture which overcomes the problem. In this we have Nx K interfaces so that it provides large number of interfaces to the data packets. Which we can save the energy in the networks. Merge networks are of different types depends up on the savage we use the network. In this 48 x 48 merge network.

IJESAT | Jan-Feb 2012

Available online @ http://www.ijesat.org 37

R.S.SINDHUTHEJA* et al. [IJESAT] INTERNATIONAL JOURNAL OF ENGINEERING SCIENCE & ADVANCED TECHNOLOGY

ISSN: 22503676
Volume - 2, Special Issue - 1, 37 40

Fig 2a: Logical Operation of a Selector

Fig 1: Modified Switch Architecture The above figure1 is the modified switch architecture and it contains N-port switches connected to one and host of the interface of some other switch. K is the number of switch input interface that are power on. The traffic from the N-input links is arranged together and then fed in to K-interfaces allowing N-K interfaces to sleep and K should a function of load and will change dynamically. Fig 2b: Constructing a 4X4 merge using Selectors To understand the above architecture let us a take small example i.e., N=4 and K=2. Here we label the input links using the letters and the input interface to the switch using numbers. So, input lines (a, b, c, d) are routed to input switch interface (1,2). In this the data packet may enter in to any one of the input lines a or b or c or d and can be routed to 1 or 2 and this choice is dynamic. Suppose if more number of data packets arrives the key idea of behind traffic aggregation is that the number of interfaces K is changed based on load like 4 x 1 4 x 4. In order to enable this form of nondeterministic routing in the N K mesh, we require a special hardware element that we call the selector whose functional behavior is shown in Figure 2(a). There are two incoming lines and two outgoing ones. When there is just one packet arrival on either incoming line, the packet is sent out on the solid outgoing line as shown in Figure 2(a). However, if two packets arrive at the two inputs, the earlier one is sent along the solid outgoing line while the later one is sent out along the dotted line. If the links have slotted transmissions and packets arrive simultaneously, then we assume some static ordering (say based on incoming line numbers) to determine which packet gets sent out over which outgoing line. To illustrate how the selector can help us build a merge network, consider the 4 K example from previously. As before, label the interfaces 1, 2, 3, 4 and the corresponding incoming lines are labeled a, b, c, d. Figure 2(b) shows a merge network that supports all possible merge combinations. The solid lines emanating from each selector is the default outputwhile the dotted line is the deflection route. To configure the network to perform a 41 merge, we simply shut off interfaces 24. Thus, packets arriving at those interfaces are dropped andonly one packet, the one that is routed to interface 1, makes it through. Similarly, to implement a 43 merge, we simply shut off interface 4 and thus three packets make it to interfaces 13 while the fourth, if there is a fourth, will be sent to interface 4 and dropped. The general structure of anNK merge network is a simple generalization of the one shown in Figure 2(b). A logN depth binary tree made up of selectors gives us anN1 merge. Next, take all the N1 deflected outputs (dotted line in Figure 2(a))of all the selectors and form a binary tree with those to get a(N 1)X1 Merge. This process iterates to create a complete merge network. An important question that arises is that of the optimality of the merge network as well as measuring its complexity. The

IJESAT | Jan-Feb 2012

Available online @ http://www.ijesat.org 38

R.S.SINDHUTHEJA* et al. [IJESAT] INTERNATIONAL JOURNAL OF ENGINEERING SCIENCE & ADVANCED TECHNOLOGY complexity of the merge network can be specific by two numbers the depth of the network and the total number of selectors used. For the network shown in Figure 2(b), the depth is 4 while the number of selectors used is 6. Indeed, we can prove that the minimum depth of anN N merge network is log2 N + N 2 and the number of selectors needed is N (N 1)/2.

ISSN: 22503676
Volume - 2, Special Issue - 1, 37 40

packet. It willgo to BUSY state if starts of a new packet at interface i-1. HIBENATE: interface is scheduled to go to hibernate when it gets a signal i.e. is data packet is coming just it takes a little amount of energy to AWAKE the interfaces.

3. SECTION III
In this section we used figure3 i.e. Interface State Diagram to give a detail explanation of how Sleep and Awake algorithm will work to conserve the energy form the networks.in this we mainly have six sates the new state is the Hibernate state which is an important state in it. When the data packets are not coming it will go into sleep state but in this even the interfaces will awake which takes a lot of energy. So hibernate helps to take the state from sleep to Hibernate when the data packet arise it will awake It take little amount of energy which is less than sleep. They are six states i.e. IDLE: interface is awake but idle. It will go to BUSY state when it begins receiving packets. It will go to SLEEP state when the interface i 1 becomes idle. BUSY: interface is awake processing packet. It will goback to IDLE state when it finishes processing the lastpacket. It will go to ABOUT TO SLEEP state if interfacei 1 done processing packet. SLEEP: interface is asleep, and if a packet comes whenan interface is in SLEEP state, the packet is dropped. Itwill go to ABOUT TO WAKE state if interface i 1becomes busy. ABOUT TO AWAKE: interface is scheduled to wake upbut it takes to be fully awake. It will go to IDLE statewhen the time is up. Or it will go to SLEEP stateinterface i 1 becomes idle before time is up. ABOUT TO SLEEP: interface is scheduled to go to sleepbut it is in the middle of processing a packet. It will go toSLEEP state when it finishes process the Fig 3: Interface State Diagram 4. SECTION IV

In this paper the newly implemented idea is the hibernate state in order to conserve the energy just it is same as the pc technology to save the energy we use to keep our system into sleep or hibernate to save the energy. Here also in LANs in order to conserve energy sleep and Hibernate state will be there the most research work is on the Hibernate. The main difference is that in sleep even the data packets are not arise the interfaces will be awake which takes a lot of energy when compared to hibernate .where as in the hibernate if the data packet is not receive it takes Sleep into Hibernate which is the Most advantages in conserving energy. The main function of it is just it take a little amount of energy to awake the interface which is less than sleep. From this hibernate we can save the energy is up to 2-3% which is more in the networks in the LANs. 5. CONCLUSION

Finally this paper gives an idea on Hibernate Algorithm which increases Power Consumption to 2-3% in Wireless LAN Networks, not only that by using various Architectures and Algorithms like Merge, Sleep and Awake also we can conserve energy in addition to Hibernate we obtain 92-93%

IJESAT | Jan-Feb 2012

Available online @ http://www.ijesat.org 39

R.S.SINDHUTHEJA* et al. [IJESAT] INTERNATIONAL JOURNAL OF ENGINEERING SCIENCE & ADVANCED TECHNOLOGY energy saving in all Algorithms and zero Packet loss.Further scope of this Paper is we can obtain pure 100% energy saving by implementing various architectures in Networks.

ISSN: 22503676
Volume - 2, Special Issue - 1, 37 40

REFERENCES
[1] M. Allman, K. Christensen, B. Nordman, and V. Paxson. Enabling An energy-efficient future internet through selectively connected end Systems. In Proceedings USENIX HotNets07, November 2007. [2] G. Ananthanarayanan and R. Katz. Greening the switch. In Proceedings USENIX HotPower08, San Diego, CA, December 7 2008. [3] G. Appenzeller, I. Keslassy, and N. McKeown. Sizing router buffers. In Proceedings ACM SIGCOMM04, Portland, OR, August 2004. [4] J. Chabarek, J. Sommers, P. Bar ford, C. Estan, D. Tsiang, and S. Wright. Power awareness in network design and routing. In Proceedings IEEE INFOCOM08, Phoenix, AZ, April 2008. [5] C. Gunaratne, K. Christensen, S. Suen, and B. Nordman. Reducing the energy consumption of Ethernet with adaptive link rate (alr). IEEE Transactions on Computers, 57(4):448 461, April 2008. [6] M. Gupta, S. Grover, and S. Singh. A feasibility study for power Management inLAN switches. In Proceedings IEEE ICNP04, Berlin, Germany, October 5-8 2004. BIOGRAPHIES

This is T.V.Vamsi Krishna studying B.Tech(ECM) in KLCE in the year 20082012.

This S.Jhansi studying B.Tech(ECM) in KLCE in the year 2008-2012.

IJESAT | Jan-Feb 2012

Available online @ http://www.ijesat.org 40