Dr. Jawid Ahmad Baktash, Abdul Basir Osmany, Mursal Dawodi
Faculty of Computer Science, Kabul University
As in today’s businesses the process of acquiring, validating, storing and protecting the data is a critical and vital subject. Therefore, effective data management acts as an important part of organizations IT systems that run business applications. There are different model for effective data management. As a big model we can refer cloud computing paradigm. Cloud computing model provides many benefits including lower computing cost, flexibility, reliability, availability, universal access, fault tolerance and virtualization. Many organizations are using could for data management based on their needs. With recent development of Internet of Things (IoT) devices, users of computing world desire connections that provide quality of service with no latency and high level of data security, which might not be obtainable through cloud computing. The same scenario runs in Afghanistan, owing to poor internet connection and low level quality of service and security, the demands for data management through the technology of cloud is not attainable. Therefore, in the year of 2012, Cisco presented a new computing model, called fog computing, as a solution for gaps of cloud computing by carrying back a part of the computation from the cloud to the edge or near to edge devices and these fog devices interact with cloud server to provide better quality of service (QoS) to the end users. With this evolution, fog computing model would be the future solution for processing and management of data in secure ways, providing quality of service (no latency) in Afghan organizations. In this paper, we first overview cloud computing model in Afghanistan including its advantages and weaknesses for data management. Next, we will overview on fog computing and its suitable architecture for Afghanistan. Finally, we explain the advantages and challenges of fog computing implementation in Afghanistan.
Keywords: Cloud Computing, Fog computing, Data Management, Afghanistan.
N the present age of information technology, existence of devices and data in everywhere is important for extending computing services and it makes our daily life easier. Therefore, data is the key product, and keeping more data typically produces more significance in today’s business world. Furthermore, with the growth of business data, the data management process also becomes a critical subject. For this reason many organizations in Afghanistan are using a computing model known as ‘Cloud Computing’, in which, resources as storage and processing power are provided to the users remotely via internet. This model provides large amount capacity for data storage and high processing power for users over the internet. With implementations of cloud computing model most data that need storage, analysis, and decision making are sent to the data centers in the cloud . Cloud computing provides infrastructure, platform and software as services to the organizations and users by which the most of issues with data management in Afghanistan overcame. As requirements for speed of data and capacity increases, and big amount of data is generated, the management process including movement of data to the cloud, instant availability, backups, recovery, confidentiality and quality of service are important and challengeable subjects particularly with unsecure and poor internet connections and bandwidth constraints in Afghanistan. Therefore the distant cloud would not be able to fully satisfy mentioned requirements specially the low latency and security of data.
In order to overcome these challenges in Afghanistan, the concept and model of fog computing would be a solution, which mainly consists of doing the data transition between edge devices and the cloud servers. The term ‘Fog computing’ for the first time proposed by researchers from Cisco Systems . Fog computing is a distributed computing model
where the edge devices of network, such as routers, gateways etc., interplay with cloud servers in order to provide services. Bonomi et al  defined fog computing technology as – “Fog Computing is a highly virtualized platform that provides computing, storage, and networking services between end devices and traditional Cloud Computing Data Centers,
typically, but not exclusively located at the edge of network”.
This paradigm concentrated on bringing computing resources as close as possible to the source of generated data to decrease latency and usage of bandwidth. In other words, Fog computing means running smaller amount processes in the cloud and moving large amount of processes near to the edge of network. This will reduce the amount of data traffic of long distance communications between clients and servers.
Fog computing model allows network operators to decrease network loads and delays and improve the responsiveness of the services by placing computing system closer to data generators and consumers.
The implementation of Fog computing in addition to Cloud computing will cause many benefits for data management in Afghanistan as; lesser consumption of network bandwidth which results higher performance and minimized latency, lower operational costs, enhanced security, better privacy, and more. Nevertheless, implementation of this platform will has its own drawbacks related to the technology platform and environment of Afghanistan which involves; resource limitation, connectivity problems, physical security issues and more efforts needed.
This research will cover one Meta question as; how to overcome the challenges against data management in Afghanistan? And three sub questions as; what specifications/requirements are desirable for data management in today’s market of Afghanistan? Does the cloud provides effective ways for data management in Afghanistan?
In definition of Fog computing a research states; “Fog computing is a geographically distributed computing architecture with a resource pool consists of one or more ubiquitously connected heterogeneous devices (including edge devices) at the edge of network and not exclusively seamlessly backed by cloud services, to collaboratively provide elastic computation, storage and communication (and many other new services and tasks) in isolated environments to a large scale of clients in proximity” .
K.Saranya1, Jaya prakash.J2, Sethu Balaji.K.G3, indicated that Fog computing is as a bridge between remote data centers and Internet of Things (IoT) devices. The author states that, this model provides a wide range of benefits such as reduced latency, decreased bandwidth and high level of security. While it also faces many security and privacy threats. They also propose a mechanism to improve security of IoT devices. The paper also states that many organizations are using cloud storage to store large amount of data. While in over all data stored in cloud is less secure. As the usage is increasing, the vulnerabilities are also increasing. However, when data in cloud are accessible through internet connection, itself causes new information security challenges. 
In another paper the researcher states; while cloud computing is broadly used but it still has some limitations. The major drawback is the connectivity between cloud and the end devices. The requirement for internet connectivity is not suitable for a large number of cloud-based applications such as the latency-sensitive ones. Fog computing paradigm presented to overcome these challenges. With the help of fog computing, processing of some data (time-sensitive) should be occurred at edge of the network, whereas others (delay-tolerant) data can happen in the cloud .
A research related to data management in Fog computing states; most of Cloud-related data storage structures can be applied or can be slightly altered to meet the requirements of Fog Computing. The main shortcoming of cloud is the high latency when dealing with edge devices. Fog Computing has arisen to solve this issue and offers Cloud-like services with low latency. To handle big amount of data produced by edge devices, which causes high network traffic, there are needs for improved algorithms to support data storage distribution mechanisms. 
A Quick Overview of Data Management
As digital market emerges, data management becomes the backbone of a business strategy. In today’s environment, it’s normal to have large amount of data including customers, employees and other business related information which should be used in decision-making, efficient business continuity and delivery of better services. Therefore, having clean, high quality and reliable data that strongly helps obtaining organizational goals is essential. To ensure these measures, there must be efficient data management solutions that could provide the following data management criteria:
- Ensuring data quality (Clean data)
- Access of data in simple ways
- Reduced data redundancies (data duplications)
- Ensuring security of data
- Existence of strong recovery plan
Overview of cloud computing in Afghanistan
It is approximately around two decades that cloud computing is using. Cloud Computing is the combination of technology and platform that delivers hosting and storage service over the Internet . This paradigm provides many benefits in processing, analysis and storage of data as efficiency, cost-benefit and more. As we have involved, a growing number of organizations and businesses in Afghanistan are using benefits of the cloud-computing model to run their businesses efficiently, providing effective services to their customers, and to increase their profit margins.
In overall Cloud computing provides three services as: infrastructure, platform, and software as services (IaaS, PaaS, and SaaS).
Platform as a service (PaaS)
Platform as a service provides a cloud-based environment in which all the requirements needed to support the lifecycle of developing web-based (cloud) applications are available. The efficiency is that you can build applications without the cost needed to buy equipment, complexity needed to manage the underlying hardware and software and hosting requirements. Organizations in Afghanistan are using this cloud utility for developing and deploying different business applications and produce them to the market quicker.
Infrastructure as a service (IaaS)
This service is providing computing resources for organizations and businesses including servers, storage, data-center space and networking on a pay-per-use basis. Many organizations in Afghanistan as telecom companies, banks and technical service providers are using benefits of this utility mainly for storing their business data because there is no need for investment on infrastructure. Cloud has a flexible infrastructure and which is scaling up and down based on requirements to support demands dynamically.
Software as a service (SaaS)
Software as a service or in other word is Cloud-based applications that are running on distant infrastructure (in the cloud) owned and organized by a third party service provider. The connection between users and infrastructure is provided by internet usually using a web-browser. Using this service, businesses in Afghanistan provide applications to their users or customers for which they can easily sign in and start to use them. These applications are accessible and usable from any connected supportive device to the internet.
Usages of cloud for data management in Afghanistan
Managing of data in cloud model might be best practice. A big number of businesses in Afghanistan are using cloud computing as their data management platform.
Several big and small businesses are using cloud platform to store and access files, using web interfaces. These interfaces are usually simple. Storage management of data done by virtualized pools of storage hosted by third party service provider based on customer specification requirements. This technology provides benefits of availability and scalability at any time and place. This way of utilizing storage services is based on paying for the amount of cloud storage without concerning about control of daily maintenance of the related storage infrastructure.
Backing up critical data continuously is a complex and time-consuming procedure for many businesses. The regular backup process contains maintaining a set of tapes or drives, organizing them and dispatch for backup purpose. There is also the chance of problems that might happen during the backup process. Therefore backing up data in this method is not safe to many issues such as running out of backup media or disk failures. Whereas in cloud based backups, businesses can send data to any location in automated based with the assurance that there is not problems of availability and capacity. The feature also provides better management of IT resources, better data safeguards and reduced backup costs.
Many organizations as banks are using cloud advantage to store their critical information for disaster recovery plan. Cloud-based service offer rapid data recovery for different emergency cases, from natural disasters to power outages. A big advantage of this service is providing lower cost and faster recovery from a network of different physical locations.
Even though the cloud offers many benefits for different organizations in Afghanistan, nevertheless owing to many technological and environmental limitations, use of this technology make some challenges for the customers who are using it.
Challenges against Cloud Computing in Afghanistan
While the cloud service providers are trying to implement the best security standards and industry certifications, storing and managing data and important files by third party providers always are open to different types of threats. The expectation of cloud users is that cloud service providers have to provide safeguard for underlying hardware infrastructure and manage them in all deployments. But, the user identity and access management is a main challenge, and should be carefully considered all the risk situations. This issue is with majority of organizations in Afghanistan which are not implementing standard information security structures and strategies. In cloud computing paradigm, everything is online, which reasons high level of vulnerabilities. Even in best practices organizations suffer from attacks and security breaks from time to time. Cloud computing is a public service due to this reason, it’s easy to run without have a complete knowledge of how to use it. In overall, cloud providers are not going to see your administration and technical skills prior granting you an account access: the whole process will take to get started with a valid credit card.
Poor internet connection
While, the only way to have access to cloud computing services is through an internet connection. Therefore, in Afghanistan this is where the biggest obstacle is happening, particularly in remote areas that do not have fair internet connection option. With the low speed internet connection in this country, it’s a big challenge with many organizations to use different cloud utilities.
As mentioned before, the only way to access cloud services is an internet connection. In some situations where the internet connection is down due to different issues, the businesses will not be able to achieve there desirable goals using technology.
how to OVERCOME DRAWBACKS of cloud computing in Afghanistan for data management?
Cloud computing model offers several benefits in data management that can help business growth, although there are leakages with this technology, it is because there is no complete perfect system. Today’s industry is beginning to adopt the Internet of Things and there is growing needs for data to be accessed and processed more rapidly and securely. Here is when the concept of Fog computing comes to operate.
Therefor we study Fog computing and its suitable architecture with advantages for data management in Afghanistan.
Cisco Systems initially presented the term “Fog Computing” as a new model to simplify and improve data transmission among distributed devices over an area in the model of Internet of Things (IoT) network . Fog computing model bridges the gap between the cloud and end devices by enabling computing, storage, networking, and data management on network nodes within the close area of IoT devices. In a simple term Fog computing is, the Cloud that is near to the network of user, means Cloud Computing that works closer to the end users is recognized as Fog Computing. Therefore, computation processes and data management not only happen in the cloud, but also happen along the Cloud path when the data traverses to the cloud. In this structure, Fog nodes should be placed closer to IoT source nodes, the result will reduce latency and save the network bandwidth compared to traditional cloud computing model and also can provide QoS especially with poor internet connections. Fog has a distributed infrastructure in which certain application processes or services are managed near to the edge of the network. But the remaining processes and storages are still managed in the cloud. It plays an important role in reducing network congestion and facilitating location awareness, mobility support, real-time interactions, scalability and interoperability . This model reduces the amount of processing and data to be delivered to the Cloud. As previous definitions Fog computing is:
- Carrying processes and memory resources on the Edge of network.
- Executing most basic process of data at those Edge resources
- Sending the results to cloud
Fog layered architecture
Fog computing has arisen recently, therefore still a standard architecture does not exist for this computing model. As researchers offered many types of Fog computing architectures like: hierarchical, network-based and layered architecture. In this study we will target layered architecture. Layered architecture is the best way to represent Fog
computing. Many studies have done to quantify the
layer-based concept of Fog architecture . This architecture would be the best suitable implementation of Fog computing in Afghanistan.
As Arkian et al.  offered Fog architecture that has four layers:
- Data consumer layer
- Data generator layer
- Fog computing layer
- Cloud computing layer
Data consumer layer is the layer in which a big number of users considered from individuals to enterprises. These users can submit their requests to three subsequent layers and acquire desired response for services.
In Data generator layer all IoT devices exist, they transmit the data to the cloud computing layer through the Fog computing layer.
Fog computing layer perform pre-processing of data. This layer provides services for context aware and low latency applications.
The cloud computing layer performs centralized control, receiving aggregated results from Fog layer and a big number of monitoring services. This layer also provides complex data analysis and final storages.
Table 1. A comparison of Cloud and Fog Computing
Advantages of Fog Computing in Afghanistan
Using the model of Fog to Cloud, actually provides two major benefits which include, advantages of both cloud and the fog models. It provides high computing and storage capabilities of cloud while minimizing traffic load and communication latencies and increasing security in Fog. Some major advantages of this model are listed as below:
Faster access of real-time data
Fog computing is a decentralized computing model, means it does not rely on a central component as cloud computing . According to another definition, Fog is a distributed computing method which focuses on enabling applications that require low latency services . This model is able to overcome the high latency problem of the cloud computing by using idle resources of devices that are near to users. There are two main advantages, minimized communication latency and avoiding two times data transferring through the same path which cause bandwidth utilization (in implementation of cloud with centralized system firstly data moved to the cloud for processing and storing purposes and then move back to the network).
delay in Transmission of data to cloud
When data is collected at Fog layer, in most cases there is no need to send them immediately to the cloud, therefore the transmission of data to the cloud can be delayed without any loss of performance. It will provide some advantages as: if there was requirements for data aggregation, it can be implemented to decrease the size of data. Also data transmission could be in timely manner to reduce the load of network traffic during the peak hours.
Quality of Service (QoS)
QoS is an important subject of many businesses of today. Madsen et al.  suggested that the “availability of different methodologies and algorithms work with the reliability of network connectivity and information, to ensure accuracy, which is crucial for building Fog computing-based projects”. While, Fog Computing is the nearest to the edge of network, therefore this model should provide new services to minimize communication latencies and the result it will deliver higher QoS. There are four scales included for QoS as connectivity, delay, reliability and capacity which are overviewed as below:
Connectivity of fog networking structure should be maintained because fog network make available opportunities through vary connectivity methods. Some Fog features as network relaying, partitioning and clustering in a diversified fog network, make the possibility of expanding connectivity.
This feature is based on latency aware applications which are important software architectural requirement in fog computing to provide the facility of a real-time streaming and processing response. Spatiotemporal system architecture  proposed a time-sensitive event processing, based on fog to mitigate high latency data transmission.
It is one of the important characteristics for secure transmission of data at the network. This feature improved by periodically checking to retransmit the data after failure.
Capacity for QoS categorized into two types, network bandwidth and storage capacity. These two characteristics are achievable with knowing data location within the fog network, it means for better connectivity data locality plays an important role.
As a result, according to the Afghanistan technological requirements and limitations major measures of QoS should be achievable through advantages provided by Fog computing model.
Security and Privacy
Based on current researches, fog computing is predicted to be the main backbone of IoT devices in the future implementations. For this reason the security measures for this model will be a critical subject. The major benefit is that Fog nodes can be protected by the same security measures (procedures and policies) that are using in a network environment. There are six important security features that should be considered in Fog computing implementations, they are as:
A. Access Control
As Fog computing structure’s a getaway among cloud and IoT devices, it transfers field data from IoT to the cloud and message from cloud to the IoT. In this situation, there is no way to access cloud databases from Fog devices. Therefore, the option of implementing access control measures is available.
For this security measure, the best solution is to configure multi layered authentication methodology in both Fog and Cloud. Therefore, any threat will only have effect on limited number of IoT devices. But, if all authentication operations fit in Fog devices, it would be vulnerable for all of the network.
Data availability means ensuring accessibility of data always and everywhere needed. In this term because Fog has position in network of IoT in which, there would be no effect to the cloud side even if communications are blocked in any of fog nodes.
Confidentiality states to protecting data from unauthorized access. In other words it means only authorized people have access to sensitive information. Fog computing supports best security measures to provide confidentiality of data which has sensible impact on IoT devices network, and a slight effect on the cloud.
It is the maintenance and assurance of information accuracy and consistency as it travels from source to destination. In Fog computing implementation on the integrity of messages, the expectation is a minimum of effect on data integrity, if end-to-end encryption is not employed and no effect on data integrity with employed encryption.
In this model, network administration team have the ability of analyzing sensitive data locally instead of sending it to the cloud for analysis. Therefore, they can keep track of data and control related devices which collect, analyze and store information. In overall, the fog computing layer should be protected by advanced security measures. For example a proxy server with a firewall/IDS. A big number of attacks should be prevented by the firewall. Unsafe intrusion attempts should be detected by IDS and prevented at the fog layer before reaching to the cloud servers. For improving security level, it’s better to keep the private and critical data on the edge of network, though sending just necessary data to the cloud.
Potential security attacks against Fog computing with possible solutions
There are some possible attacks while running Fog model; we overview them with provided solutions to mitigate the risks, they are as following:
A. Virtualization attacks
To mitigate these type of attacks Multi-factor Authentication Intrusion Detection System, Role-Based Access Controls, Process isolation techniques and user data isolation methods should be implemented.
B. Data security
Solutions might be implementation of advanced encryption algorithms, security key management, network monitoring tools and classification of data.
C. Communication attacks
Mitigations should be limiting the number of connections which are allowed, multi factor authentication mechanisms, isolating compromised nodes in some cases, executing communication encryption and finally implementation of transport layer security measures.
D. Web security attacks
The solutions can be consistent software updates, securing codes and patches, using and updating anti-virus protection system and auditing in timely manner firewall system.
E. Wireless attacks
To deal with this types of attacks these security measures are essential: advanced authentication system, communication encryption, securing wireless routing and wireless security protocols.
Fog Computing issues in Afghanistan
While Fog computing provides many benefits in scope of data management. On the other hand, this paradigm has different drawbacks due to wireless connectivity, decentralized management, and power failure . According to Afghanistan the major failures would be:
Comparing Fog devices with the Cloud, there is less resources available in Fog computing in processing and computation. Yet, Fog have the ability to executed data on devices by making Fog clusters as the server to deal with huge data processing using commodity machines. 
In Fog environment most of device are connected by wireless connections and the number of these connected devices increase over time. This would be a challengeable subject in some cases due to interference. It will affect the overall bandwidth. Furthermore, In Afghanistan the environment structure caused line-of-sight which also should raise as an issue. To address these issues, it is important to use models and technologies having flexibility with the number of connected devices within Fog to overcome the connectivity limitation.
Alongside different information security issues in Afghanistan there is a big challenge against information technology overall, which is physical security. It also include Fog computing, different organizations might not accept the risks to invest on technology infrastructure for Fog computing implementation within this country.
This paper has shown that businesses and organizations in Afghanistan need the most suitable data management paradigm. There are different usages of cloud computing for data management in in this country. But, according to environment and technology limitations in Afghanistan, the cloud model is not able to deliver efficient services needed by different organizations. Therefore, this paper focused on a new model named Fog computing. This model as an extension of cloud computing will provide desirable services for businesses in this country. Fog computing will offer the advantages of low latency service, high level of security and privacy, data quality of service and more. Although the model provides many benefits, it also has its own shortcomings in this country such as wireless connectivity, resource limitations and physical security treats.
REFERENCES. Ravandi Babak, Papapanagiotou Ioannis, “A self-learning scheduling in cloud software defined block storage”, IEEE 10th International Conference on Cloud Computing, IEEE, 2017.
. Flavio Bonomi, Rodolfo Milito, Jiang Zhu, Sateesh Addepalli, “Fog computing and its role in the internet of things”, in Proceedings of the first edition of the MCC workshop on Mobile cloud computing. ACM, 2012.. Yi, Shanhe, Li, Cheng., & Li, Qun, “A Survey of Fog Computing: Concepts, Applications and Issues”. In Proceedings of the 2015 Workshop on Mobile Big Data. New York, NY, USA: ACM, 2015. . K.Saranya, Jaya Prakash, Sethu Balaji.K, “Survey on Fog computing and its application benefits in real life”, International Research Journal of Engineering and Technology (IRJET), Volume: 05 Issue: 05, May-2018, ISSN: 2395-0056. . Carla Mouradian, DialaNaboulsi, Sami Yangui, Roch H. Glitho, Monique J. Morrow, and Paul A. Polakos, “A Comprehensive Survey on Fog Computing: State-of-the-art and Research Challenges”, IEEE Communications Surveys & Tutorials, 2018. . Vasileios Moysiadis, Panagiotis Sarigiannidis, and Ioannis Moscholios, “Towards Distributed Data Management in Fog Computing”, Wiley, 2018, Article ID 7597686. . Harjit Singh Lamba and Gurdev Singh, “Cloud Computing-Future Framework for e-management of NGO’s‖”, IJoAT, July 2011, Vol 2, No 3, Department Of Computer Science, Eternal University, Baru Sahib, HP, India, ISSN 0976- 4860. . John K Zao, Tchin-Tze Gan, Chun-Kai You, Cheng-En Chung, Yu-Te Wang, Sergio José Rodríguez Méndez, Tim Mullen, Chieh Yu, Christian Kothe, Ching-Teng Hsiao, San-Liang Chu, Ce-Kuen Shieh, Tzyy-Ping Jung, “Pervasive brain monitoring and data sharing based on
multi-tier distributed computing and linked data technology,” Frontiers
in Human Neuroscience, vol.8, no. 370, 2014.
. Redowan Mahmud, Ramamohanarao Kotagiri and Rajkumar Buyya, “Fog computing: A taxonomy, survey and future directions”. In Internet of Everything, Springer, Berlin, Germany, 2018.. Amir Vahid Dastjerdi, Harshit Gupta, Rodrigo N. Calheiros, Soumya K. Ghosh, and Rajkumar Buyya,“Fog computing: Principles, architectures, and applications,” a chapter of Internet of Things: Principle & Paradigms, Morgan Kaufmann, Ed., USA, 2016. . Arkian Hamid Reza, Diyanat Abolfazl, Pourkhalili Atefe, “MIST: Fog-based data analytics scheme with cost-efficient resource provisioning for IoT crowdsensing applications”, Journal of Network and Computer Applications. 2017. . Jianhua Li, Tiehua Zhang, Jiong Jin, Yingying Yang, Dong Yuan and Longxiang Gao, “Latency estimation for fog-based internet of things,” 27th International Telecommunication Networks and Applications Conference (ITNAC 2017). IEEE, 2017. . Henrik Madsen, Bernard Burtschy, Grigore Albeanu, and Florin Popentiu-Vladicescu, “Reliability in the utility computing era: Towards reliable fog computing,” in Proceeding of the 20th International Conference on Systems, Signals and Image Processing (IWSSIP), IEEE, 2013. . Kirak Hong, David Lillethun, Umakishore Ramachandran, Beate Ottenwälder, and Boris Koldehofe, “Opportunistic spatio-temporal event processing for mobile situation awareness”, 7th ACM International Conference on Distributed Event-Based Systems, USA, July 2013. . Redowan Mahmud, Fernando Luiz Koch, and Rajkumar Buyya, “Cloud-fog interoperability in iot-enabled healthcare solutions”, 19th International Conference on Distributed Computing and Networking, USA ACM, 2018. . Ranesh Kumar Naha, Saurabh Garg and Andrew Chan, “Fog Computing Architecture: Survey and Challenges”, researchgate, 22 November 2018, ISBN: 9781785616372.