RFID IMPLICATION WITH CLOUD

This research paper is written by me in my university days, i am sharing it out with other peoples, please share your views in comments. Asad Malik

---

1.      ABSTRACT:

Radio Frequency Identification (RFID) uses radio waves to track the movement of products. The item is identified by a unique number that is passed on easily to a pc. Middle size companies are experiencing RFID execution limitations.  The limitations are poor processing power, limited computational resources; poor data management etc., because of these issues RFID is not able to deliver what it is capable of, hence businesses are finding it difficult to adopt it. In this paper I have proposed a cloud platform for RFID system which will help RFID solution providers to meet RFID limitations and provide solutions which can meet the requirements of the businesses. Our platform will help meet the challenges which are un-answered yet and are affecting the reputation of radio frequency identification. Our platform will also help small and mid-size companies to achieve the full advantages associated with radio frequency identification. Our platform also provide data analysis and processing features which in traditional computing are very expensive and mid-size companies are not able to process and generate trends from the gathered data, but in our proposed platform the cost of data analysis and processing is reduced drastically. 

2.       Introduction:

Radio Frequency Identification (RFID) [1] is growing as a significant technological innovation used for monitoring products and resources all over the globe. It can help medical centers identify expensive devices more quickly to enhance individual care, drug companies to reduce counterfeiting and logistics suppliers to enhance the management of portable resources [2]. It also guarantees to enable new usefulness in the supply chain management system by monitoring goods from the manufacturing place through to the retail outlet or point of sale [3]. As a result of the potential benefits of RFID, many of the globe's major suppliers have implemented RFID labeling for pallets and cases delivered into their sale centers. The effect of this RFID activity in the retail store industry is likely to effect on around 200,000 producers and providers worldwide, and will boost the market for software and hardware to support RFID. Extensive use of RFID labels increases the likelihood of people being monitored though individual belongings. The effects are stunning - buys can be connected to the bank cards that were used to make particular buys, which allows hyperlinks to particular promotion based on individual investing styles. It is often described as a life enhancing technological innovation with regards to its potential impact on business techniques. However, in many ways it is a sensible major step on from the bar code technology as a way of getting enhanced work performance through automated way. When used together with other technologies it can sense things to determine their recognition monitor their place and recognize features such as pressure and warm range.

RFID has surely changed the way of tracking in very less amount of time. In current era it has influenced our daily life style, it is embedded in goods which we buy from market. In USA government is considering a suggestion to embed an RFID tag in body of newly born baby for identification purposes. Nowadays, this technological innovation is implemented worldwide for various programs and simultaneously on the other hand radio frequency identification has few issues associated with it. These issues are effecting the growth of radio frequency identification and are stopping its deployment in organizations.

The main issue which is affecting RFID is that retail organization thinks that acquiring technology is the total expense of implementing radio frequency identification in their company, but in reality the 80% of the cost in implementing RFID is that of software architecture and this is the main issue which is forcing mid-size organizations against RFID.

With my proposed cloud platform for radio frequency identification, small and mid-size organization will not have to worry about software architecture, they just have to buy hardware and the rest will be handled by our proposed cloud platform. As it will be offered to them as software as a service so there will big difference of cost and complexity level will be drastically decreased due to our proposed solution.

Cloud computing is referred to a network of remote computers available over a network typically the internet. It is further divided into 3 major types: Software as a service (SAAS), Platform as a service (PAAS), and Infrastructure as a service (IAAS). In software as a service software is offered as a service over the internet, the user is charged on pay-per-usage model. In platform as a service platform for deployment of application is offered over the internet as a service in which user don’t have to undergo the complex and time taking activities of managing servers etc. In infrastructure as service hardware equipment i.e. storage, processor etc. are offered as service over the internet.

3.      Literature Review:

Given how bug this issue of providing a strong architecture and platform is that other researchers has also researched on it, the closest work to ours is researched by Melanie R. Rieback and co-authors. In their paper has tried to address the security problem associated with RFID. They have proposed a platform which handles key management and authentication [7]. They worked on the idea of centralizing management of RFID tags. The difference in their proposed system is that they are just implementing this system using traditional server and traditional way of computing, where our proposed platform is cloud based and it acquires all the advantages of cloud computing as well, similarly they are just going to centralize the authentication and management whereas our proposed platform starts from deployment and handles all the issues associated with RFID like handling large volume of data, processing hundreds of request at the same time, performing data analysis on run time and allowing some technical users to write their own application using our platform.

Christian Floerkemeier and Matthias Lampe in their paper has discussed few reasons which causes RFID to fail in Ubiquitous computing applications [5]. The authors have stated that due to weak I.T infrastructure and platform if multiple cards are scanned simultaneously the system failed to detect some cards despite of the fact that card is in the range of the reader. They have discussed some other reasons as well of system not been able to read cards but they are related to physical aspect of cards. There is not any solution provided by the author that how we should address those limitations, only authors have described the issues faced by RFID execution in certain environments. Author has identified an issue associated with RFID that it needs to be addressed in near future. The issues stated by the author are successfully addressed by our proposed platform.

Fusheng Wang and Peiya Liu has stated the issues associated with data management of RFID system [6]. They have stated that RFID system produces bulk of data and the data produced by RFID system is time dependent, dynamically changing and it is actually a big challenge to manage all of the data. Similarly the author have pressed the issue that there should be such strong platform which can address the issue of data management in RFID.  Further explaining the point author have said that the big infrastructure which is needed to address data management issue should also be capable of running such methods automatically which can analyze all of the data  and can produce such results and trends which can help the businesses financially. Authors have proposed a system which filters RFID data and can be used in different RFID applications across the world. Using traditional computing, handling large volume of data was always an issue and it is also identified by the author in this paper,  our proposed system is quite different and modern in term of technology as it uses cloud data bases to meet the issues produced by bulk data.

Several existing systems are discussed above and the difference is also very clear in our proposed system and the others, as none of the system is using cloud platform for managing data and deployment of rfid systems.

4.      Motivation:

An RFID tag is so small that it can be attached anywhere, once attached it emits radio frequency signals and a reader can identify the tag if it is in its range. This tag can be attached to computer software as every tag has unique serial code stored in it and further data analysis can be performed on the data produced by RFID system.

While I was working on my final year project which used RFID for identification purposes, there need for such platform was raised. There were 2 sides of our system one being the server side and one client side, on server side we were supposed to manage all data processing i.e. implementing queues to process and load data in database. We soon realized that using traditional computing architecture will not serve our purpose so we decided to develop a cloud based platform with expanded scope like it was supposed to manage all the data processing and analyzing issues, with addition to this we also moved deployment related work on our platform.

After our proposed platform all of the difficulty in deployment and setting up I.T infrastructure for RFID is sucked out. Now it has become more easy and is “in-range” for small and mid-size companies to use RFID system for their business and achieve the advantages associated with RFID to boost their business without getting involve in the difficulties of I.T infrastructure. 

5.      Discussion:

Nowadays, RFID is used across the globe to identify items and to monitor the movement of goods. Many applications are developed using RFID which are successfully serving the purpose, but the fact is that there are few issues and limitations which are hurting the growth of radio frequency identification technology and therefore because of these limitations, organizations are not able to achieve the full advantages associated with radio frequency technology and hence this thing is taking those organizations away from this wonderful technology.

For example one big issue organizations encounter is that there I.T infrastructure is not strong enough to serve simultaneous requests consistently and as a result the data is lost. This loss of data cause huge financial deficit. Take an example of a store where there are cartons of juices and each carton has RFID chip embedded in it. Now store keeper has to scan all of the cartons to include them in stock but after store keeper has scanned all items we came to know that 4 requests were not processed completely and were discarded so now our inventory will show 4 less cartons.

Another problem which is faced by most of the companies / organizations using radio frequency identification is that RFID systems produces huge amount of data that is difficult to manage. There was a research done by a Chinese RFID manufacturing company and they came out with the result that RFID produces 10-100 times more data as compared to bar code systems.

The issue of deployment has always been there with RFID from very beginning. To setup RFID hardware is not an issue but to setup the software side of RFID which can interact with our server is a big issue. As there are hundred different of RFID vendors out there in the market and each have their own set of APIS, so it becomes difficult to write code back and forth for new hardware which can then meet the requirement to connect with server which In most of the cases is placed somewhere else and is accessed remotely.

Our proposed platform will be based on software as service architecture. Software as a service now a days is becoming very popular because of its distribution model [10]. It is also very useful In terms of cost because user only pay for the usage of server [11]; it is not like traditional computing where you have to pay for the whole month despite of the fact that you use it or not. It is the biggest advantage because at the end of the day what matters is the cost and software as a service has dropped the cost very much and these services are now available to user at very low cost. It is because of this software distribution model that now customers are charged at per CPU cycle, so one fact of cost is eliminated from our proposed platform. Now because of the cost reduction it means that our platform can be accessed and used by mid-sized and small companies and now they can also access all the advantages associated with RFID and cloud computing. For example in case of data analysis the servers an technology is very expensive and those mid-size companies and small companies cannot buy servers and I.T infrastructure to process and analyze data, but using our platform it is very easy to process and analyze data. In current era the importance of data is very much and business takes their financial decisions after data analysis, trends and patters found after data analysis are very helpful in financial aspect as it help businesses to grow and prosper. So due to cost reduction, now every size of business whether it is small or large can now take full advantage of all the features associated with cloud computing [12] and radio frequency identification.

The purpose of this paper is to suggest a cloud platform which can address the issue stated above and can interact with RFID to meet the challenges imposed on RFID.

a.      RFID Implementation barriers:

Scalable I.T infrastructure:

One main issue which is associated with RFID even after years of its execution is of scalable I.T infrastructure to manage and process the requests and to avoid data loss. The main advantage associated with RFID is that it can process multiple tags at single time and that advantage is not fully acquired if we don’t have a scalable platform which can handle all those requests. Consider a fact where we don’t have such platform and our system has read 100 tags for us in a second but now our system is not able to process all those requests and suddenly data is lost which results in financial loss.

Saygin, C in his paper "Adaptive inventory management using RFID data" [8] has stated that loos of production data results in loss of profits. In his paper the author has studied and presented the models for inventory management and has presented data the heart for any RFID system. Data in current era is sometimes referred to as money, by processing data for analyzing trends and then taking decisions on those processed data, businesses can increase their profit.

Many believes that there should be such mechanism where all the requests can be handled properly and data loss should be 0%. Industry leaders believes that it might not be possible to process all the requests simultaneously at a time but there should be a mechanism where we can at least put all the requests in queue and later can be processed sequentially.

Currently there is no such mechanism established which can guarantee that there will be no data loss in situations where RFID reader can read over 200 tags in parallel. So big organizations are not taking any risk and are not deploying RFID in their businesses.

It is very sad to see that big organizations like IBM etc. are not paying any attention towards this and are not trying to propose any solution publically for this issue which is hurting the RFID technology very badly. 

Huge volume of data:

Another issue is that RFID system produces huge volume of data that is very difficult to manage and process. If we compare the data produced by RFID system with that of data produced by bar code system, RFID system produces 10-100 times more data than bar code system.

Another thing is that RFID data is dynamic and it is not of the same type i.e. the bar code only reads the tag of items where as there are hundreds of use cases for RFID like product tracking monitoring, loyalty cards etc. so the data produced from RFID system is not of the same type and kind as it is in bar code system. So to store and process data we need some scalable platform to perform this operation.

RFID deployment Issue:

Consider the fact that a retail store has RFID implemented in their different branches and suddenly after 4 years of time they decide to open a new branch and for this branch they have to buy a new RFID reader because the old RFID readers which are working in their old branches are no more available, now as every RFID vendor has their own set of APIS now client has to ask some company to re write the whole communication mechanism for them, it is a big issue and not yet answered. 

6.      Proposed cloud platform:

Our proposed platform is a cloud based application, which will be offered as a service to our customers. It will be based on pay-per-usage model, customer will only pay for what they use. Our platform will have multiple features like data management and processing, deployment section and application section.

In cloud computing the big advantage which we achieve is that of distributed processing and vertical scaling. The cloud computing architecture is very elastic and the server’s configuration in cloud can be extended as the demand increases. For data processing we have got the edge from cloud computing which offers distributed computing and the data can be processed very fast and very effectively across multiple nodes. The data is highly available in our proposed platform because of cloud computing vertical scaling and backing up feature. If one node is down the second node is up in no time and the data is available again.

We have proposed queue system for our platform for processing requests and not allowing application to handle the request directly. By following this method we are making sure that none of the request is dropped out and all the requests are processed. In our platform if a request is failed to be processed it is added back to the end of queue and is not dropped out of the system, so it will come again and will be processed again this process will continue until the request is not completely processed and is stored in our database. It is one of the brightest feature of our platform because it is one of  the biggest challenge imposed to cloud computing in real world. In traditional computing the request is dropped out if it is failed to be processed there is no such system which can process it again another limitation which we have addressed in our proposed platform using queue system is that there is a limitation in number of requests which can be processed at a time,  not more than 150-200 requests can be addressed simultaneously but using our platform this limitation is no more there and unlimited number of requests can be processed in parallel. Using distributed processing we analyze our data for shopping trends and customer can then take their business decisions based on the analysis done by our platform.

The other section which will be in our application is of application section which will present a GUI for all the services which are offered by our proposed platform. It will contain the information regarding data, analysis results and other information associated with our platform.

The third challenge which we have tried to address in our proposed platform is related to deployment. Our proposed system will have a deployment section where customer will be asked to select the vendor and model of their RFID reader, after passing these parameters customer will be prompt to download a jar file which will be installed in customer PC , the rest will be handle by that jar file and our platform. The jar file will pass the request to our cloud platform and further processing will be done by our platform. It will simplify and change the whole concept of deployment, in traditional way the deployment is done by deploying the whole application on client end but in our case I think we have revolutionized the way of deployment, following our new way of deployment client only need to have a maximum 1 MB jar file on their system and the rest all of the work is done by that jar file.

So the proposed platform is very easy to use and the complexity level is very low, client has to make sure that computer has a stable internet connection and has installed the jar file the rest of the of the complication of processing the request is handled by jar file and our platform.

The other advantage which is associated with our proposed platform is that mid-size and small level companies now enjoy the data analysis and processing facilitation whose cost in real world with traditional computing is very high, and these companies are not that financially strong to buy those machines and servers, but using our platform its cost is reduced and they will just have to pay for what they use. For example if they are using our server for data analysis foe 4 hours then they will have to pay only for 4 hours not for the whole month where I traditional computing companies were forced to pay for the whole month or they were supposed to purchase packages which were for month despite of the fact that they want it only for few hours.

One advantage which comes with our platform is that of performance. Our platform based on cloud architecture has high availability time as compared to any other computing architecture. Clients don’t have to be worried about availability and also don’t have to be worried about availability of application. Cloud architecture is very flexible and elastic and fault avoidance can be done in run time as cloud whole architecture is based on virtualization and to move application and data to another instance is not a big issue. The cloud computing actually means a network of computers connected together so it actually by default spread the application across multiple servers [9] because it works on the concept that moving computation is cheaper then moving data and application, so if an instance is down or a node is down simply the next node in network replaces that node and the system is up again and down time of server is very less almost un-noticed.

Meanwhile I want to list few of the advantages that can be achieved by using cloud computing

·         Cloud computing helps to keep the price of application down. Users who are using that particular application can access it from anywhere without taking care of their pc, laptop, phone etc. All they need is internet connection to get access to their application deployed on cloud.

·         Cloud computing is very much scalable as compared to our traditional computing architecture. Its biggest advantage is i

·         Applications deployed on cloud computing are highly available as it follows “Horizontal scaling” architecture and the data and application are spread across multiple servers having mirror images of the applications. So that if one server is down the other server is made available which serve the users until that server is back live.

So as you can see there are many benefits associated with cloud computing, I have listed only major ones but there are hundreds and thousands of advantages which are associated with cloud computing.

a.      Solution for Scalable I.T infrastructure:

As stated above in implementation barrier section that it is one of the issues associated with cloud computing and there is need of a scalable platform which address this challenge.

Our proposed platform has the ability to meet this challenge, as it is a cloud based platform and our platform will be offered on software as a service model, it has the ability to process more than 200 requests simultaneously. A cloud server specification can be increase with workload increasing so it doesn’t affect our platform if requests are 200 or more.

Secondly our proposed platform has queues implemented to process the requests so data loss in now 0 %, because we are not dealing with data directly all the requests which are received to our platform are added in the queue and our platform process those requests sequentially.

b.      Solution for managing high volume of data:

As discussed in implementation barrier section one of the issues is that RFID system produces huge volume of data which is very hard to maintain using traditional computing. As data is not just for storage but we also have to process and analyze data and then this data is used to make business decisions.

In our traditional systems, we only have a networked database hosted on a single server linked with our application and the data is stored directly into the db. There is no such mechanism to take backups and the data is never highly available in our traditional computing. But in cloud computing with our proposed platform data is replicated across multiple servers and is processed through the queues before going into the database. We are using the cloud architecture with the concept of vertical scaling where our data will be spread on multiple servers and will always be highly available in case of if any server is down. The data is made available because of our backup servers where data is simultaneously stored after processing.

With the use of cloud computing the big advantage which can be achieved is that of distributed computing. In our proposed platform we are relying on distributed processing to process our data and get the useful trends out of our data after processing. The other advantage is that If we do all the stuff of backing up and distributed processing using our traditional computing then the cost of the system will be increased by 500% as compared to cloud computing where the cost is very low and you pay only for what you use.

In general services offered by cloud computing are general and elastic and are virtualized. It reduces the cost of platform, hardware drastically. Another advantage which comes is that all of the services of cloud are available via internet and the advantage of stability is always there in cloud computing. In current era there is no other computing architecture which is more scalable then cloud computing which forces us to use cloud computing architecture to meet this challenge of managing high volume of data.

c.       Solution for deployment of RFID:

As I discussed earlier the issues associated with cloud computing deployment and I gave an example of a retail store having multiple branches. The solution which we propose to this is that our cloud platform which will have jar files for certain vendors and customer after selecting their vendor will be able to download those jars.

The jar file will have set of APIs and functions; functions will only pass the request received from the reader and the rest will be handled by our platform which will further the process the request.

In our proposed solution the deployment over head is no more there and secondly customer doesn’t need to put very huge systems on client end. All they need is a set of jar file downloaded from our platform, a computer with normal 256 kb/s internet connection and a RFID reader which reads the card. All this can be setup in between $100 - $200, so it is a very cost effective solution.

7.     Conclusion:

In this paper I proposed a cloud platform for our RFID system which addresses most of the challenges faced by RFID, we effectively used cloud computing architecture to address these issues. The biggest beneficiaries of our platform will be small and mid-size companies because they can now use a powerful platform to fully achieve all the advantages associated with RFID system and can enjoy all the luxuries in very low cost. I suspect that another solution which addresses these issues without involving cloud computing will be very costly and in fact very hard to implement.

With our platform customer will just have to signup, select a package which suits their needs and then they will just have to download a jar for their system, the rest will be handled by our platform. So with our proposed platform all the difficulty is sucked out from client end.

8.        REFRENCES

[1] – Shepard, Steven. RFID: radio frequency identification. New York: McGraw-Hill, 2005. [2] – Najera, Pablo, Javier Lopez, and Rodrigo Roman. “Real-time location and inpatient care systems based on passive RFID.” Journal of Network and Computer Applications 34.3 (2011): 980-989. [3] – Attaran, Mohsen. “RFID: an enabler of supply chain operations.” Supply Chain Management: An International Journal 12.4 (2007): 249-257. [4] – Armbrust, Michael, Armando Fox, Rean Griffith, Anthony D. Joseph, Randy Katz, Andy Konwinski, Gunho Lee et al. “A view of cloud computing.” Communications of the ACM 53, no. 4 (2010): 50-58. [5] – Floerkemeier, Christian, and Matthias Lampe. “Issues with RFID usage in ubiquitous computing applications.” Pervasive Computing. Springer Berlin Heidelberg, 2004. 188-193. [6] – Wang, Fusheng, and Peiya Liu. “Temporal management of RFID data.” Proceedings of the 31st international conference on Very large data bases. VLDB Endowment, 2005. [7] -  Rieback, Melanie R., et al. “A platform for RFID security and privacy administration.” USENIX LISA. 2006. [8] - Saygin, C. "Adaptive inventory management using RFID data." The International Journal of Advanced Manufacturing Technology 32.9-10 (2007): 1045-1051. [9] - Zhang, Qi, Lu Cheng, and Raouf Boutaba. "Cloud computing: state-of-the-art and research challenges." Journal of Internet Services and Applications 1.1 (2010): 7-18. [10] - Buxmann, Peter, Thomas Hess, and Sonja Lehmann. "Software as a Service." Wirtschaftsinformatik 50.6 (2008): 500-503. [11] - Sun, Wei, et al. "Software as a service: An integration perspective." Service-oriented computing–ICSOC 2007. Springer Berlin Heidelberg, 2007. 558-569 [12] - Armbrust, Michael, et al. "A view of cloud computing." Communications of the ACM 53.4 (2010): 50.