 and welcome to today's lecture on storage technology. We have already discussed about cache memory, main memory which are used as memory devices of a computer system. And if you look at the hierarchical memory organization, particularly the virtual memory system that you have discussed in the last couple of lectures, you will find that secondary storage is the last component in that hierarchical memory organization. And storage technology provides that secondary storage and so our discussion will not be complete without considering storage technology. We shall discuss about different types of devices which are used in secondary storage like magnetic disks, optical disks, magnetic tapes, flash memories and nowadays solid state disks. And in another lecture, we shall discuss about how we can improve the reliability by discussing multiple disks together and different red levels, red level 0 to 66. And today let us primarily focus on the different types of technologies used to build secondary storage. So, magnetic disks were introduced back in 1965 and perhaps surprisingly they are still ubiquitous hard drives. So, if you look at any computer system, you will find hard disk is always present and that is become ubiquitous with the computer systems. And this is the bread and butter of storage in desktop and server systems. So, without magnetic disks, no computer system can be built or we cannot see nowadays and magnetic disks are used for long term storage, non-volatile storage for data, non-volatile means when even when the power is removed information will be retained. And this has already provided the additional level of memory hierarchy through virtual memory system that we have already discussed. So, these are the different features of magnetic disks. Number one as I said it is non-volatile, disk storage is non-volatile and it consists of a collection of platters. So, platters means different disks you can say with two surfaces and each platter has magnetic material on both sides. So, you can say this is a platter circular disk and magnetic material is deposited on both sides. As a consequence, the total number of recordable surfaces is 2 into number of platters. So, number of platters that you are having will decide how much memory you can how much information you can store instruction or data whatever it may be. And the platter size is I mean the diameter is in the range of 3 to 4 inches smaller sizes are also available. But this is the common size 3 to 4 inches normally 3.5 inches is the diameter of the magnetic disks and bigger sizes can be used for performance and smaller for cost. So, smaller the diameter it will be cheaper and larger the diameter it will provide you better performance. And the rotational speed is in the range of 3.6 K that means 36000 revolutions per minute to 15000 revolutions per minute and actually the performance is gradually improving. And this is a magnetic disk as you can see you have read write heads these are the read write heads and each of them is representing a platter and as you can see each plate is having two sides that means 1 and 2 corresponds to two sides of platter 1, 3 and 4 corresponds to two sides of platter 2, 5 and 6 corresponds to the two sides of platter 3 and so on. Of course, the last platter you cannot really access the bottom side that is why only 11 surface is shown 12th surface is not accessible. So, this is how I mean magnetic disk is organized you have got a number of such platters providing you 2 into number of platters that is the number of surfaces available. And the way information is written on magnetic disk is shown in this particular diagram this is that disk surface which rotates at a particular speed which you have seen it rotates and it is that disk is placed just under a read write head. So, this is that read write head there is a coil that is an electromagnet that means if current passes through it will be magnetized in a particular direction and because of that magnetization the iron particles will get oriented in a particular way depending on the current flowing in that electromagnet direction of the current flowing in that electromagnet. For example, before the disk surface comes under the comes under the electromagnet you can see the iron particles are half a jadley distributed there is no orientation of the iron particles. However, as it passes through the read write head the iron particles gets oriented in a particular way depending on the current direction. So, this is the current direction now so bit 0 is written and disk is moving in that direction. Similarly, bit 1 is written when current passes in the opposite direction so in this particular case current is passing in the opposite direction so bit 1 is written. So, in this way as the medium rotates the head can writes data on the surface it can write bit 1 bit 0 and so on one after the other on the disk surface reading can be done in a similar way. In such a case what happens the as the disk moves under the I mean already written data moves under the surface it induces current in this coil and depending on the direction of the current in this coil you can read 0 or 1. So, this is how reading and writing takes place on the magnetic surface with the help of a read write coil by placing it on top of the magnetic surface. So, one point you should note at this young here at this point you can see the read write head is coming in physical contact with the magnetic surface as a consequence there is wear and tear. So, there will be wear and tear because that magnetic surface is coming in close contact here for example, this is the electromagnet. So, it will become in physical contact as a consequence the magnetic material will get eroded with time that is the reason why the life of a magnetic disk is not very long. So, this is how reading and writing takes place there is a concept called formatting to make the read the disk I mean to for storing information on the disk a technique known as formatting is used that means before a magnetic disk is used for storing data it is formatted and formatting is a process that maps the disk surface and determines how data can be written. So, this surface is earlier it is a I mean it is a without any demarcation about where you can store data which part has to be kept free, but as it is formatted then it identifies on which particular area you can store data and in which particular area you have to keep blank and you have to keep gap. So, during formatting each platter is divided into tracks and around the disk surface. So, you will find there are tracks so it will be this is one track a circular ring this is another track. So, in this way up to certain I mean certain distance up closer to the centre the writing can be done. So, in this way this is the track 0 and this will be the track say n. So, you can have n tracks n plus 1 tracks if it is n minus 1 there will be n tracks on the surface and the number of tracks is increasing with the advancement of technology as you can see it can vary from 10 k to 15 k tracks per surface. On a single surface you can have 10,000 to 50,000 tracks so quite a large number of surface I mean tracks you can have on the magnetic surface actually that depends on the resolution of the redried head smaller it is it can store on a smaller area of the surface. And each track is divided into sectors so suppose this is one track and this track is divided into a number of sectors so this is one sector this is another sector this is another sector this is another sector. So, in this way you have a number of logical sectors so you are dividing into a number of sectors and the total number of sectors per track can vary from 100 to 500 so you can have 100 to 500 sectors per track and a sector is the smallest unit that can be accessed that means whenever you are reading data from disk or writing data into the disk so the minimum unit that you can access is a sector that means a sector can be read or more than one sector can be read so multiple of sectors can be read and on a single sector you can have 512 bytes to 4 k bytes of data per sector so this is the typical value the number of bytes that you can store on a sector varies from 512 bytes to 4 k bytes. So, this gives you some information about a disk and here a formatted disk is shown you can see this is a disk surface shown and as I said that outer track is the 0th track so this is the 0th track and 0th track is divided into a number of sectors 0, 1, 2, 3 so in this particular example there are 18 sectors per track and the number of sectors on each track is same irrespective of whether it is closer to the center or that means it is same in the 0th track and the nth track that means number of sectors per track is identical so you can see various sectors on different tracks are shown so this is how logically a disk surface is I mean used and where you can store information so this part the central part is used for holding this disk on the disk system because you have to firmly hold it so that it can be rotated at a particular speed and there are four areas defined at the time of formatting that operating system creates four areas on the surface number one area is known as the boot sector so it stores the master boot record a small program that runs when you first start the computer that means normally you know whenever you put a disk and start the computer the boot sector is read and boot sector is used for the purpose of I mean bringing the system up second is your file allocation table so that is a log that records each files location and each sector status that means so far as the so far a programmer is concerned he is concerned about files he stored different files and these files are to be mapped to different tracks and different sectors of a track that mapping is done that information is stored in this file allocation table or fat then third area is root folder it enables the user to store data on the disk in a logical way and finally you have got data area it is the portion of the disk that is that actually holds data so you can say those boot sector file allocation table root folder these are overhead only area where information data can be stored is the data area as you can see in this diagram different areas are shown so that boot sector is that green portion is the boot sector then is that file allocation table fat that red track and that file allocation table is duplicated for the purpose of I mean higher redundancy so there are two file allocation tables stored as you can see two fat copy one fat copy two then root directory that blue circle blue track is showing the so you have started in this way first the boot sector then the fat then the second copy of the fat then it goes to the second track where you are storing the root directory root directory is being stored root directory is stored so these are the components which is essentially overhead and information is stored only in this part of the disk now in the past all tracks had the same number of sectors as I have already mentioned and this made it easier to implement means for instance reading a sector on a track required the same angle of allocation so earlier what was being done as you have seen this is the track closer to the center and this is another track which is at the outmost layer of the disk so this is the disk so we have seen that you have you are dividing into a sector so for the so this is one sector this is one another sector and this is one sector you can see that length of this which is away from the center is more than the length of this part where which is closer to the center so the bit density is different in different tracks as you go away from the center the bit density is smaller compared to the bit density which is on the track which is closer to the center so that is because of this particular technique where all tracks have the same number of sectors so this meant that bit density was lower on the outer tracks than that of the inner tracks as I have mentioned now subsequently with the introduction of zone bit recording varying number of sectors per track can be used so this is the this is one innovation which was used to increase the capacity of a disk so capacity expressed as the area density measured in bits per square inch is improving so area density is essentially area density which is representative of the capacity per unit area I mean unit inch that is actually tracks per surface on a disk into bits per inch on a track so this gives you the parameter area density and actually it specifies the number of bits that you can store per unit area and this is improving over time as you can see in the starting from the year 1988 to 1996 the increase in area density was 29 percent then from 1996 to 2003 the improvement was 60 percent per year and from 2003 onwards to recent years it is 100 percent per year that means every year the area density was doubling so you can see it is even I mean better than the Moore's law however currently the rate of increase has dropped and it is only 29 percent per year that this is the rate at which the area density is increasing so you can see whenever you are using this type of innovations like zone bit recording and so on you are able to increase the capacity of a disk and the rate of improvement is quite high as you can see now how you are performing reading and writing so there is an arm with a read write head for each recorded surface so there is a separate read write head you can see this is shown here separate read write head which is available for each surface so all heads on the same track on all surfaces so what is what happens that all the heads move together so that means what happens your read write head if this is the read write head if it is pointing to track number 0 so it can move in this direction track sorry track number 0 is this one track 0 track 1 in this way it can move with the help of motor that is called stepper motor with the help of a stepper motor the read write head is positioned on different tracks of a surface and this read write head assembly you can say all the heads on different surfaces move together so if you have got say n platters you will be having a 2n read write heads all will move together on different surfaces of course you may say 2n minus 1 because the last platter which is at the bottom will not have one surface will not be available so you will be having 2n minus 1 read write heads which will be reading and writing on the surface now those set of tracks is called cylinder what is a cylinder actually the concept of cylinder is coming in this way suppose this is one platter then at the bottom is there is another platter then there is another platter then there is another platter in this way you have got large number of platters now let us consider the read write head is positioned to track number i so track number of i of this platter on both surfaces then the track number of i of the other platter on the same track so all will form a kind of cylinder so essentially you can read from this particular track of this particular platter and for this platter also you can read together so these are forming a kind of cylinder so that is why the concept of cylinder has come in because different tracks on different plates is forming a kind of cylinder so to read write a sector the discontroller sends a command that causes the arm to move over the proper track so this operation is called seek so there are different type of times you will encounter the first one is known as the seek time what is seek time seek time is the time to position a read write head on a particular track so initially you know that read write head can be arbitrarily positioned then it is then read write head is withdrawn and whenever it reaches 0th track from there it starts moving towards other tracks with the help of that stepper motor so seek time is the time required for positioning time required to position on a particular track then the seek time is one of the important performance measures for magnetic discs since it is a very important performance measure there is confusion about its definition so people will specify about minimum seek time maximum seek time average seek time so people will get confused which one you have to take because minimum seek time means the minimum amount of time if the head write is very close to the 0th track then seek time will be small if it is very close to the center then you have to take it back towards the 0th track and the seek time will be longer so that is the reason why minimum seek time maximum seek time and average seek time these are recorded these are reported the first two are easy to measure because it depends on the actual position of the read write head at a particular position so here as you can see the read write head is here so you have to move to the 0th track so this is the 0th track and this is the nth track so then it will take longer time when on the other hand if the read write head is at the 0th track then the seek time will be smaller so average seek time is computed over all possible six so you can the read write head can be at any on any one of the tracks of the end tracks and as a consequence you have to take the average of them so the seek time can vary between 5 millisecond to 15 millisecond these days nowadays the improve in spite of the improvements it is of the order of millisecond because mechanical motor is involved in the process of I mean pressing a read write head on a particular track and that is the reason why it the values I mean varying between 5 millisecond to 15 millisecond and obviously this time is a very long time for a CPU you know nowadays processors are operating at the rate of few gigahertz so you can see this 15 millisecond or even 5 millisecond is too long a time so for example for 5 gigahertz CPU 5 millisecond means 20 million clock cycles so just for positioning the track positioning the read write head on a particular track you will take 20 million clock cycles that is too long a time so an applications is lower seek time due to locality of course although this is the I mean worst case time you can say but in practice because of the principle of locality actual seek time will be much lower it has been reported that the actual seek time is about 30 percent of the specified seek time so computing seek time is not straight forward but the arm speed is not constant another very important parameter is the read write head I mean which is holding the read write head is hold by arm this is called arm that arm speed of that arm is not constant so it is speeds up and slows down because in the beginning it will be having slow speed then you know just like you know you are driving a car initially speed will be low then the speed will it will pick up then as you try to stop it it will slow down so it is a motor so the arm may stop to reduce the vibration so it may be allowed to stop without giving power so that the vibration is not present and another parameter is rotational latency so first parameter is seek time second is rotational latency what is rotational latency we have seen that seek time is the time required to position a read write head on a particular track now you have to place it on a particular sector because that that file I mean which is mapping to tracks and sectors in terms of tracks and sectors to read data you have to position not only on a track you have to position on a particular sector so on a particular track you have to bring your read write head on a particular sector so that is actually decided by the rotational latency and average rotation time is roughly half of the worst case time that is your 0.5 by 10000 rpm if you see the speed is 10000 rpm so that is roughly equal to 3 millisecond so you can see rotational latency is of the order of 3 millisecond so we may say that this is of the order of 3 millisecond and this can vary from 5 millisecond to less so it may be 5 millisecond or 4 millisecond like that now comes the another parameter that is your transfer time so the time to transfer a sector from the magnetic material to the head so that means you have position a read write head on a particular sector now you have to read it that read write head will read it and transfer it to the through that through the from the there is a controller through the controller it will send to the processor so the transfer time is the time to transfer a sector from the magnetic material to the head and it depends on the sector size disc size rotation speed width density speed of the electronics at the head and the disc controller so you can see this transfer time is dependent on a number of parameters and these parameters will decide the transfer time and typical value is in the range of 40 to 120 megabits per second so this is the speed at which the data transfer will take place after the read write head has been positioned on a particular track and then on a particular sector and after that this is the rate at which data transfer will take place from the read write head to the processor so you can say that average access time then your transfer time so there are three components and you can say that average disc access time it has got several components one is average seek time plus average rotational delay plus transfer time and there is some overhead because of the controller controller is holding some electronics and that controller will take some time to transfer the data to the CPU so this is the average disc access time comprising four components average seek time average rotational delay transfer time plus controller overhead so you have to take into account all these four components whenever you find out they try to calculate the average disc access time. So, let us consider an example to crystallize the idea let us assume that average seek time is 5 millisecond rotational species is 10000 revolutions per minute number of bytes per sector is 512 number of sectors per track is 500 transfer rate is 50 megabits per second and controller megabytes per second everything is specified in terms of bytes and controller overhead is quite small 0.01 millisecond now considering I mean with the help of this parameters you can calculate what is the time required to transfer a file from the disc to a processor so question is what is the time to read a file consisting of 2500 sectors. So, let us try to calculate this so we can calculate in this way time to read the first track so that it will vary from track to track so after it has been positioned on the first track then subsequent tracks can be easily accessed because seek time will be 0 for other tracks so that is why time to read the first track is separately considered that will consist of 5 millisecond assuming that seek time is 5 millisecond then the rotational delay is 3 millisecond then the time to read say 500 sectors we know we have seen that there are 500 sectors per track and to calculate to time required to transfer 500 sectors will be even calculate it and you will find that it is again 5 millisecond and then you have got that controller overhead which is 0.1 millisecond so this is the total time that will read the first track comprising 500 sectors and this is equal to 13.1 10 13.1 millisecond. Now you have other four tracks so time to read the second track second and subsequent tracks you can say will be equal to this will no longer be present so this will be equal to 3 millisecond plus 5 millisecond plus 0.1 millisecond so that means this will be equal to 8.1 millisecond so total time to transfer 2500 sectors will be equal to 13.1 plus 8.1 4 into 8.1 because you have to read 5 tracks so 2500 tracks I mean 2500 sectors and 500 sectors per track so you have to read 5 tracks so first track will take time required will be 13.1 millisecond remaining each track will require transfer of data from each track will require 8.1 so it will be 4 into 8.1 for the remaining 4 tracks so total time will be equal to 45.5 millisecond so this is how you can find out the access time for different situations. Now we shall focus on one very important aspect that you can exploit that is the locality of reference often an application reads consecutive sectors you see although a particular file can map to different tracks and different sectors but usually a particular file is stored in consecutive sectors of a particular track so that locality can be exploited special locality and that special locality can be exploited and the technique that is used is known as read ahead so most hard drives used uses read ahead so what is this the disk has a buffer that stores sectors after one just read that means we have seen that whenever a disk is accessed minimum unit of access is one sector normally one sector will be transferred then it will read another sector and so on but what you are doing in this particular case you are not only reading one sector you are reading another sector you are reading ahead and this can be as large as 4 megabytes and it can be that buffer can be as large as 4 megabytes where you can store subsequent sectors and it is just a cache of sectors so we have already discussed about the use of cache memory so it is acting as some kind of cache memory a buffer where you are reading some additional sectors and storing the information of those additional sectors so this can also store sectors that need to be written or to the disk. So in this here not only reading you can also do it for writing purpose that means you will be writing into the buffer then you will be transferring to the disk so both for reading and writing you can use this concept of read ahead which is akin to the concept of cache memory in the context of main memory system so transfers to and from the buffer are at times restricted by the speed of the IO bus so here you know a particular around we should discuss about that a computer system will have several buses one is CPU memory bus which is the fastest that means the CPU transfers data with the CPU another is CPU IO bus so it is like this so here let us assume you have got CPU and there is a bus and this is the CPU memory bus to which the cache memory and main memory systems are connected so this is your cache memory and this is your main memory. Now another bus you can have that is provided by with the help of a IO controller and this will provide another bus known as IO bus so this IO bus is relatively smaller I mean slower compared to the CPU memory bus because the rate of transfer of data that takes place to this IO bus is I mean is normally slower IO devices are connected through the IO bus and as you can see here the data rate can be 300 megabytes per second so compared to few gigabytes per second here the speed is 300 megabytes per second it is in the bus mode bus mode means you are transferring all the data together 4 megabytes you are sending one after the other not you are not reading one sector then another sector another sector it is already available in the buffer and from the buffer all the data is transferred in the bus mode so this is how you can exploit the locality with the help of this read ahead concept. So this gives you the trend for bus drivers I mean disk drives you can see here three parameters shown that yellow line corresponds to shipments in PB so the volume is increasing at a very high rate the second the pink line corresponds to cells in billions so you can see one very important feature I mean aspect or point you must note here you can see although the shipment is larger shipment is increasing because of the drop in price the cells in billions has remained same that means the money that is the transaction of money that is actually taking place is more or less constant over the years that means the volume is increasing price is declining actual amount that is being that involved money that is involved cells in billions of dollars that is remaining more or less constant. So for disk drives one PB corresponds to 1000 terabyte one terabyte corresponds to 1000 megabyte and one megabyte corresponds to 10 to the power 6 byte so you can see we are considering very large numbers large volume you can say which is stated in terms of PB and not in terms of terabyte or megabyte so this is the trend of disk drives and this trend has remained more or less same over the years. Now we shall switch gear and consider another type of disk so far we have considered magnetic disks now we shall focus on optical disks so many new technologies have evolved over the years particularly but they have not become successful commercially but this optical disk is the one of the challengers to magnetic disk is the optical disk which has become quite successful commercially so optical disks are quite successful an optical disk is a high capacity storage medium an optical drive uses reflected light to read data so here reading and writing is taking place with the help of light instead of putting the read write head on top of the magnetic surface here you are doing the reading and writing with the help of light as a consequence wear and tear does not exist we have seen whenever you put the read write head and top of the magnetic surface and when it rotates there is a wear and tear which does not exist whenever you do the reading and writing with the help of light so to store data the disk metal surface is covered with some tiny dents known as pits and flat spots known as lands so you have got two types of areas pits and lands which cause light to be reflected differently so you are trying to read with the help of light and depending on where it is I mean from where it is getting reflected from if it is on pits it will reflect in one way if it is on lands it will reflect in another way so when an optical drive signs light into a pit the light cannot be reflected back as we shall see within the next diagram this represents a 0 bit value and a land reflects light back to the source representing a one value so this is shown with the help of this diagram so in this particular here as you can see this is the light source laser source so which is falling on the land so which is representing one since it is falling on a flat surface it is reflecting the light and then there is a prism which drivers it which I mean there is a total internal reflection and then it goes through a sensor and that sensor converts it to light to electrical signal and you go get a one so this is how you get a one on the other hand whenever the light is falling on a pit which is representing bit 0 then what is happening the light is getting scattered in all directions so it is not getting reflected like the previous case and in the consequence there is no light present that no light will pass through the sensor so it is the absence of any signal light signal is considered at 0 this is how a 1 and 0 can be read from the optical discs so in pieces the most commonly used optical storage technology is called the compact disc or read only memory that is your CD ROM so a standard CD ROM disc can store up to 650 megabyte of data or about 70 minutes of audio nowadays because of the popularity of music and also the popularity of CD ROM these two has become very it has become very common to store music on CD ROM so that is why you know instead of megabyte you are also stating in terms of time that is the amount of audio signal or music that you can store 70 minutes of audio so once data is written into a standard CD ROM disc the data cannot be altered or overwritten so this is one characteristics of the CD ROM that means this writing this writing is cannot be modified so some kind of burning has taken place in certain areas so with the help of some special device these pits and lands are created and it is permanent that is why it is called CD ROM compact disc ROM ROM stands for read only memory you cannot do the writing so early CD ROM drives were of single speed and read data at a rate of 150 kilobits per second hard disk transfer data is about 5 to 15 megabytes per second so you can see there is a significant difference in the rate of transfer of data so I mean CD ROMs are much smaller compared to hard disk in terms of data transfer rate and CD ROM drives now can transfer data at the rate speed of about 7800 kilobits per second and data transfer speeds are getting faster and faster with time and CD ROM is typically used to store software programs so whenever you buy a software and it is permanently stored in the CD ROM and that is delivered to you on the other hand CDs can store and CDs can store audio and video data as well as text and program instructions so different types of information can be stored in CD ROMs then another variation is known as DVD ROM DVD stands for digital version video disk so DVD RAM stands for digital video disk read only memory and it is and is being used in place of CD ROM in many new pieces so nowadays this is also has become very popular and standard DVD disks can store up to 9.4 gigabytes of data enough to store an entire movie so you can store an entire movie into it and dual layer DVD disk can store up to 17 gigabytes so DVD disk can store as much data because both sides of the disk are used along with sophisticated data compression technique so in case of CD ROM we have seen the capacity is much smaller but in case of DVD we are using some sophisticated technique by which you are also doing data compression to store more data and that is the reason why the storage capacity is much higher and then another variation is there CD recordable CD recordable means you can write only once so with the help of a special device you can do the writing only once so drive allows recording personal CDs you can create your personal CDs and conventionally which you use that is CD recordable CDR but data cannot be overwritten once it is recorded on the data or to the disk so this is the most common type of disk we encounter there is another type of DVD ROM or CD you can say CD re-writeable so this variety is very costly and not very commercially successful so CD re-writeable drive allows recording of CD of a CD re-recording of a CD you can write read just like your it may consider it as CD read write CD RAM then write new data over already recorded data so this gives you shows you different varieties optical compact disk digital video disk DVDs 4.7 inch diameter so one new technique that is used here is instead of constant angular velocity optical disk uses constant linear velocity so earlier we have seen that angular velocity is small because you know the so when the read write head is here the time that is required to read for this part is same as this part so angular velocity is same but now what is being done in case of CD ROM you will be so this is the center and it is written in this manner in the form of a spiral so here since you are storing in the form of a spiral so this is one particular sector this is another sector this is another sector and so on so that is and you are reading using of constant linear velocity so you have got constant linear velocity of reading you are here time to read this part and this part is same so this is your constant linear velocity so CD RAM uses this constant linear velocity and there are different kinds of read only memory CDs are 0.6 gigabytes DVDs 4.7 gigabytes these are very common also double sided DVDs are available we have already discussed about CDR CDRW then DVDR and DVDR RAM so difference between CDR and CDRW is in case of CDR you can write only once in case of CDRW you can write many times then comes the magnetic tapes this is as old as the magnetic disc technology much more it has got much more storage space very long access time so you can see here on a it is in the form of a tape just like your audio tapes so on a single tape you are it is having a number of tracks track one track two track three track nine so you can transfer parallely a number of bits different bits so nine bits you can transfer parallely whenever you are reading from a tape and unfortunately it has got very long access time because you are reading sequentially you have to start from one end and then you have to go ahead just like your audio tapes and in magnet since it is made of magnetic material it wear outs very quickly because of physical contact of the read write head and primarily these magnetic tapes were used for archival purposes so in the early years in our computer center we had large number of magnetic tapes where all the various programs and data were to be stored in magnetic tapes and primarily they are used in archival purposes so with this let us come to the end of today's lecture so in my next lecture we shall discuss about some other alternatives because magnetic tapes are becoming increasingly I mean becoming outdated and new technologies are coming in which I shall discuss in my next lecture thank you.