CC BY
11TECHNICAL SCIENCE
Munqath Alattar,
Information Technology Research and Development Center, University of Kufa, Kufa, P.O. Box (21), Najaf
Governorate, Iraq Marwah Yaseen
Master in Computer Science - Cybersecurity, Independent Researcher, Najaf Governorate, Iraq
DOI: 10.24412/2520-6990-2022-31154-30-39 ESDB: A MODEL DESIGN OF A CLOUD COMPUTING EDUCATIONAL SCALABLE AND EFFICIENT DATABASE: ARCHITECTURE AND PROPERTIES
Abstract
Cloud computing is a developed distributed computing architecture for highly configured and scalable applications that is mainly based on the remotely delivered services, on-demand payment model on the shared resources and a third party to take care of the maintaining the services provided. Cloud database management system is one of the cloud computing services that provides a distributed database computing that is delivered on a cloud computing platform as a service where the provider maintains the physical infrastructure and database and leaves the customer to manage the database's contents and operations. This kind of services are basically constructed to serve the scalability benefits, where the system is able to extend its processing, storage and other hardware or software resources dynamically up and down according to the needs. In this paper, we focus on the educational and learning organizations that can use the services that are delivered through the network in a scalable way which can provide more advantages to the students, teaching staff, researchers, and administrators. We propose an architecture for the educational cloud computing database services named Educational Scalable Database (ESDB) which provides high scalability features and reusability.
Keywords: Cloud computing, database management systems, Scalable systems, educational systems, scalable cloud databases.
Introduction. The main ideas behind cloud computing are the remotely delivered services, on-demand payment model on the shared resources and a third party to take care of maintaining the provided services. These became the major reasons for the widespread adoption of cloud computing architecture, where this architecture involves shifting the services from the local computers and servers to the remote computers across the network, generally the Internet. These services refer to the system software, hardware infrastructure, storage, security, process units (Abadi, 2009). This makes scalability and elasticity two interesting topics to research and develop to scale up and down the cloud computing services.
Cloud computing is a developed distributed computing architecture that differs in some characteristics that makes cloud computing more applicable for the highly configured and scalable applications.
Table 1 shows the scalability, abstraction, economy, and configurability features of cloud computing over distributed computing (Ian et al., 2008).
1.1 Cloud Computing Models
Cloud computing can be structured in private, public and hybrid models, where the key difference of these models is the network management to handle the services and the client's scope.
Public: service providers offer their services to the client through the public network generally through the Internet to a wide range of clients and it is not limited to a single specific organization.
Private: Or internal clouds that are used only by a single organization within a private network, generally intranet or by virtual private network. The organization itself can manage the private cloud or it can be done by other external providers.
Table 1
CLOUD COMPUTING FEATURES OVER THE DISTRIBUTED COMPUTING.
Feature Cloud Computing Distributed Computing
Scalability Highly scalable and parallelizable workload Scalability limitations
Abstraction Can be encapsulated as a unit and delivered in a different scaling levels Not applicable as an abstracted unit of services and delivered as a product.
Economy Massively economic due to its scalability and the pay-as-you-go pricing mechanisms. High expense considerations
Ability to Configure Dynamically configured and deliver on-demand Costly and hard configuration process
Hybrid: Is a combination of public and private cloud models. Where there are parts of the services run in the private clouds and other parts run in a public cloud. This can help to get over the limitations of the two models. Table 2 shows a comparison of some benefits and usage limitations for each model of service.
Choosing a model of a cloud computing generally based on the organization work scenario. Hybrid cloud can be a better choice due to its characteristics that have both private and public advantages, but still some business needs to be totally constructed in a public schema, like educational services. Where in the educational
field of services, the client doesn't need to give a con- ganized, they need to shift the whole risks, manage-straint on how all the system components work or or- ments, maintenance and backups to the service provider
(Qi et al., 2010).
Table 2
CLOUD COMPUTING MODELS COMPARISON.
Mode Advantages Disadvantages Usage
Public Cloud • There is no need for an initial investment on infrastructure where this shifts to the service provider side. • Limited control on data by its owner and the settings of the network and security are restricted by the service provider. • Highly scalable projects, • Limit experienced staff and • Costly projects
Private Cloud • The organization have the full control over the data, network, and security settings • There should be an initial cost planning for the infrastructure and that required a highly skilled staff. • Organizations with sensitive information like e-banking, ecommerce, and the military.
Hybrid Cloud • Provide tighter control and security compared to public clouds, while still facilitating on-demand service. The design requires carefully determining the best split between public and private cloud components. Dominant type for most organizations (Qi et al., 2010).
2. CLOUD COMPUTING ARCHITECTURE Cloud computing can be classified according to the services level they provide. Services can be application software, frameworks, database management services or even a hardware shared infrastructure resource. As shown in table 3, these services can be divided into three abstractions: software, platform, and infrastructure. According to the cloud services users' requirements, they can choose the appropriate cloud service abstraction model.
Software as a Service (SaaS) is the highest level and the nearest to the user that provides the software applications to the user through the cloud.
SaaS provides a software application for a specific purpose that can be customized dynamically by the user to fit their needs and limitations, where these
Microsoft Azure, Google AppEngine, Force.com and Facebook's developer platform are example PaaS providers. The lowest level of the cloud computing services and the nearest to the provider is the Infrastructure as a Service where the computer infrastructure (such as server, CPU, memory, storage, network etc.) is provided as a service in a usage-based pricing model. The resources can be dynamically scaled up and down according to the users' needs.
Typical examples are Amazon EC2 Service and S3 (Simple Storage Service) where compute and storage infrastructures are open to public access with a utility pricing model. Eucalyptus is an open-source Cloud implementation that provides a compatible interface to Amazon's EC2 (Elastic Cloud Computing), and allows people to set up a Cloud infrastructure at premise and experiment prior to buying commercial services
Some other architectures like the one that provides the database management systems over the cloud can
applications are accessible remotely through the internet with the pay-as-you-go pricing mechanism. Salesforce.com online CRM, GoogleApps, Microsoft Dynamics CRM NeSuite, Oracle (Ian, 2008) (Sudipto, 2011) Google Docs are some commonly used services examples for the SaaS model services. This cloud computing services level can be described as a Distributed Application Provider level.
Platform as a Service is the middle stage of the services that the application deployment platform is provided as a service over the cloud. In this level of cloud computing services, the service provider offers an environment to deploy, serve, and scale the applications built by the customer. That means the provider automatically manages and provisions the customer's applications on the platform. (Sudipto, 2011).
be called Database as a Service. Also assigning some security duties to a third cloud party mentioned as Security as a Service. In DBaaS, databases and their relations are delivered in a scalable way. The database scalability can be obtained by several mechanisms like partitioning the database. The architecture that is going to be focused in this study is the database level of service where this study placed it in between the platform and the infrastructure architectures.
2.1 Database as a Service
A Cloud database management system is a service that provides distributed database computing that is close to the SaaS service level of cloud computing. It is a database that is delivered on a cloud computing platform provided as a service where the provider maintains the physical infrastructure and database and leaves the customer to manage the database's contents and operations. It is the sharing of information re-
Table 3
CLOUD COMPUTING ABSTRACTION MODELS.
Services Resource Management Example
Software as a Service (SaaS) Software application and Web Services slaseforce.com and Google docs
Platform as a Service (PaaS) Software framework and Database Google AppEngine and Microsoft Azure
Infrastructure as a Service (IaaS) Hardware resources, computation (VM) and storage block Amazon EC2 and GoGrid
sources between multiple devices over the cloud. According to the fast development of the networking technology, the total cost of the data managements over the cloud has decreased especially while the pay-as-you-go mechanism is possible (Yvette and Sunguk, 2012) As a result, putting the responsibilities of database management tasks to a third parties has a growing interest for much lower cost and more scalability.
Cloud DBMS may utilize all the components of the stand-alone DBMS or may add more features like combining one or more elements (like combining data structure and query). That leads many organizations to think about shifting to the cloud model.
3. CLOUD COMPUTING DATABASE PERFORMANCE FACTORS
The basic features of the cloud database management systems is characterized by the transactional processing termed as ACID that stands for Atomicity, Consistency, Isolation and Durability (Amr, 2012).
Atomicity which is also described as the all-or-none property of each transaction. It refers to the ability of the DBMS to guarantee that either all of the tasks of a transaction are performed or none of them are. In atomicity, if a part of a database query transaction fails, then the entire query fails, and vice versa.
Consistency means that the result of each transaction will be seen by everyone who reads from the database, and they have to see the latest version of that data (Jaroslav, 2013) This cloud computing database feature ensures that the database remains in a consistent state although the transaction succeeded or failed. For example, database systems in a bank system always must give correct data. A weak consistency feature provides a faster system than a strong consistency one (Anandhi and Chitra, 2014).
Isolation means that every operation is isolated from any effect of other operations. Each operation cannot access the data of another operation that is in running state and has not finished yet (Anandhi and Chitra, 2014).
Durability states the persistence of the transaction affects if the transaction is committed although if any failures occur later. That means when the operation is declared as a success status, it will persist and survive from the system failure.
3.1 CAP BASE Properties
In contrast with ACID properties, consider now the triple of requirements including consistency (C), availability (A) and partitioning tolerance (P), shortly CAP:
Consistency means that whenever data is written, everyone who reads from the database will always see the latest version of the data. The notion is different from that one used in ACID.
Availability means that we always can expect that each operation terminates in an intended response. High availability usually is accomplished through large numbers of physical servers acting as a single database through data sharing between various database nodes and replications.
Partition Tolerance means that the database still can be read from and written to when parts of it are completely inaccessible. Situations that would cause this appear, e.g., when the network link between a significant number of database nodes is interrupted. Partition tolerance can be achieved by mechanisms whereby writes destined for unreachable nodes are sent to nodes that are still accessible. Then, when the failed nodes come back, they receive the writes they missed.
4. DATABASE SCALABILITY
Scalability in its general term means the system ability to extend its processing, storage and other hardware or software resources dynamically up and down according to the needs. Where a scalable system is the system that its performance can be improved after adding additional hardware, infrastructure, or software resources to it (Divyakant et al., 2012).
Database scalability is the most important factor that the cloud platform should provide to process a massive amount of data in the cloud computing environment and used to improve the performance of a database (Jaroslav 2013).
The scalability property can be obtained typically through two different ways, vertically (also known as Scale Up) or horizontally (also known as Scale Out). Vertical scalability means adding more resources to a single unit of the system to increase the throughput, like addition of processors, memory units or hard drives. This kind of scalability enables more database servers running to handle more workload on the system. Horizontal scalability means adding more units to a system and treating them as a single unit like adding new computers to a distributed system. It is demonstrated by distributing the data and load of simple operations over many servers with no resources shared among those servers. This kind of scalability enables the system to scale from one-to-many web servers (Divyakant 2011) (Anandhi and Chitra, 2012).
Choosing scalability depends on the system work nature, where vertical scalability can be a good choice when the time factor is critical, and the scalability should be done as soon as possible. But it costs more as the size of the application increases. While horizontal scaling can be considered as a cost-effective approach, specially it can be done using the existing solutions to scale. On the other hand, horizontal scalability is not always a good choice for a low budget where it requires the application to be constructed again to work in a distributed way as a single unit and that adds an additional cost (Anandhi and Chitra, 2012). Number of application users also can be considered as a factor to choose the suitable scaling way. For example, if the number of the users is approximately fixed, vertical scaling is wise, but if the users' number is extremely changing, vertically scaling can be a more expensive choice than horizontally scaling.
The database query should be constructed in a scalable computing way to achieve the scalability optimization on the scalable cloud computing system.
4.1 Scalability Architectures
Any system that planned to use the cloud computing scalability should be constructed in a scalable way. Scalability can be obtained in several methods like separating the database to several pieces in the server or spreading the database pieces to several server machines which is called sharding. Choosing a scalability mechanism is based on the project nature and there is
DATABASE
no one method that can fit all projects. Database scalability is classified into two main methods: horizontal scalability (scale out / in) and vertical scalability (scale up / down). Horizontal scaling includes adding more units of resources and treating them as a single unit while vertical scaling includes adding more resources within the same unit that means increasing a single unit capacity. Table 4 shows the main compared characteristics of the two scaling mechanisms.
Table 4
PROACHES COMPARISON
Feature Horizontal Scalability Vertical Scalability
The Technique Separating a data to multiple tables by putting different rows in a different table Creating a table with fewer columns and using additional tables to store more columns
Usage Used for more complex systems that require high scalability features which can't be reached by the vertical scalability. Used for the less scalable, easy to implement and fast response systems
Implementing Scaling out across multiple database servers Scaling up to more powerful CPUs and servers
Complexity More complex, where it basically based on the data distributing over multiple tables and that means design the system to be distributed Less Complex to implement that it just needs to add more tables for the extra columns and less distribution processing.
Advantage There will be no single unit of failure and easily scalable by using the old low performance equipment in a single high-performance system Low expensive scalability technique to implement and used by a non-distributed application.
Disadvantage The programming and management complexity and existing applications may need to be redesigned again to tend to the distributed computing model. Required a downtime while new resources are added
In Horizontal scalability, when the database scaling reaches the hardware capacity, the database should shard across servers and replicate the data. The replication process makes the system slower.
To choose the suite scaling method, the following questions should be asked:
• Which design can address the requirements?
• What characteristics are important to the system, is it the request and response time? or, is it the fault tolerance? If so, till what degree?
• How much data will be stored?
• What is the system workload? for example, the number of users and what type of access they have (read / write).
• What is the rate of the data growth?
• What are the types of the data that is going to be stored?
• What will be the nature of the system queries (point queries, range queries or scanning the entire relation)?
• Does the system include sharding, replication or clustering?
The scalability often has some partitioning processes, data partitioning also can be horizontally and
vertically. In addition to the scalability, data partitioning offers more benefits such as:
• performance improvement
• availability that mostly there will not be a single unit of failure
• operational flexibility like management, monitoring, backup, and data restore, data administration and cost.
• match the data store with the data type and usage pattern by using a Blob data store for the binary data and a document database for the structured data.
This study proposed an architecture and educational environment that provides services that are horizontally scalable. Horizontal scalability generally means partitioning the data over multiple database tables and adding more data in an additional data table. This partitioning can be obtained by three main methods: Range Partitioning, Hash partitioning and List partitioning.
Range partitioning: In range partitioning, each set of data is assigned to a partition according to a value range that uses the "from - to" concept to define the range. Each partition has a higher value that should not be exceeded which is called a Partition Key. For example, the data with values less than 500 is stored in partition 1 and the data with values more than or equal to
staff of an educational organization. Table 5 shows a brief comparison of the three partitioning techniques.
To add more scalability enhancements to the database system, some partitioning operations can be added like partitioning advisor operation. This operation can suggest a partitioning action to the administrator in any case of data workload. In addition, an interval partitioning can be set to automatically partition the database when the data exceeds a specific threshold.
This differentiation of the horizontal scaling strategy leads to keep some considerations before choosing one of them such as:
• Queries across multiple partitions are more time consuming than on a single partition
• Replication can reduce the separate search in different partitions
• Executing separate queries for each partition and joining the data over the application level can reduce the multiple partitions joins queries.
Table 5
DATA PARTITIONING TECHNIQUES._
Strategy Advantages Disadvantages
List Partitioning Group and organize unordered and unrelated sets of data in i natural way. Hard to sequentially scan of entire relation
Hash Partitioning Full Sequential scan and searching only one partition for the point queries. Hard to answer range queries
Range Partitioning Full Sequential scan and searching only a few partitions for the range queries. Hard to organize unrelated sets of data
500 is stored in partition 2. It is easy to implement the range type of partitioning and so flexible to work with. But from the other side, rebalancing is difficult to obtain by using this partitioning method. Rebalancing over the partitions means designing the partition key again or even re-partitioning.
Hash Partitioning: When there is no clear partitioning key to define, a hash algorithm can be designed to equally distribute rows among partitions by giving partitions approximately the same size. It is also an easy-to-use alternative to range partitioning especially when the data has no range pattern to conclude a partitioning key.
List Partitioning: It enables the control of how the rows map to the partitions and that by specifying a list of separated values for each partition. For example, partition 1 contains the data related to the students and partition 2 contains the data that is related to the faculty
• It is important to define a map for the query to locate the correct partition. That is because all the partitions in the horizontal scalability are almost the same structure.
• Rebalancing is the most critical point in the horizontal partitioning.
• Monitor the partitions number growth.
In the next section, a brief description of the cloud computing services and their application in the educational field is given. After that, some general existing cases of the educational organizations and universities that use cloud computing in their academic services are explained. Furthermore, we summarized the main cost Benefits of cloud computing in Education. Finally, we propose an architecture for the educational cloud computing database services named Educational Scalable Database (ESDB) which provides high scalability features and reusability.
5. CLOUD COMPUTING IN EDUCATION
The educational and learning organizations can also use the services that are delivered through the network which can provide more advantages to the students, teaching staff, researchers, and administrators. Each organization uses different levels and scope of the
services according to their limitations, which is mainly the cost limitation. The move of the educational organization toward the cloud services made the cloud services providers to pay more attention on developing an educational cloud service, which can be delivered in a different cloud computing models where the students, lecturers, administrative staff, researchers and developers can interact with the system in different models according to their kind of services and their policies (see Figure 1).
Also, it can be provided in any cloud architecture or even it can be delivered as Education as a Service (EaaS). Some of the universities that already applied the cloud computing to offer their academic services and gain a vast advantage result, for example:
• University of California decreases the software licenses costs by shifting the ownership of the software to a cloud computing service provider in a payment effective way. Also they reduce the IT staff from 15 to 3 (Deka and Malaya, 2012).
• Florida Atlantic University has trimmed the IT cost by at least 600,000$ by virtualizing the data center using Hyper-V Deka and Malaya, 2012).
Administrative _§aa§
Staff
IaaS
Developers —► PaaS
Teaching Staff, Students and Researchers
Fig. 1. The Cloud Educational Entities Interaction levels.
• Aga Khan University in Pakistan found that moving to use the academic cloud services is more secure, viruses protecting and reduces the calls to the IT department to 60%.
• Pike County School has replaced over 1400 workstations with IBM cloud to reduce the cost.
• Singapure Polytechnic University shifted to use the academic cloud services for the cost saving, energy efficiency and the dynamic scalability benefits.
And many other universities choose the cloud services like Tacome Public School that uses Azure IoT and Portland Public School and Vanderbuilt University which use the Live@edu Email Services. The cloud-based e-learning provides a lot of benefits like online courses access and online communication with the professors. The practical advantages of the academic cloud computing services can be briefly listed as:
• Access from anywhere and anytime.
• Cost saving where no software licenses (free Software or PAY as you GO)
• share content easily and collaborative Learning
• auto backup and monitoring
• more support for teaching and learning
• high availability
• Opening to various universities and advanced research
• green technologies
• Increased functional capabilities
• Offline usage and synchronization opportunities
• Increased openness of students to new technologies
• Critical information can be kept secured by a cloud security provider
Along with these benefits, there are many limitations on applying the cloud computing in an educational field, like:
• Not all application run on cloud
• Organizational support and standards
• Security and protection of sensitive data that although cloud computing offers a sufficient security service, it is still not safe to keep the sensitive information online in the cloud.
• Lack of confidence can be considered as a limitation of using such services
• Internet and connection speed (Saju, 2012).
These advantages and limitations are classified
based on the academic level of use delivered for students, teaching staff, administrators, researchers, and the developers. Table 6 shows the advantages and the limitations of using the cloud computing services for these academic entities (Shashi, 2013).
5.1 Educational cloud computing systems
Organizations like Microsoft, Google and Amazon are providing grants and free access for universities, colleges, researchers, and students to be used with less effort for enabling facilities for their academic activities (Ramkumar, 2013). Amazon, Google, and Microsoft are the mostly requested cloud computing services providers. They offer a lot of cloud services for educational purposes, Table 7 illustrates the main services, usage, and limitations for each offered service.
5.2 Cost Benefits of cloud computing in Education
The main cost advantage of cloud computing in
the education field is the total cost of ownership (TCO) versus the virtualization where the cost is per user per time of use. Along with that, cloud computing offers a fast queries search, no local backups needed for the online data and resources can be found easily. An example of such systems is distance education based on cloud computing and Virtual Computing Lab (VCL).
6. EDUCATIONAL SCALABLE DATABASE (ESDB) ARCHITECTURE
The database management system for the cloud computing can be illustrated in three main levels: application level where the user integrated with the database with a query, database level that contain the data, relations and the tables, and finally the storage level where the operations of backups, encryption and monitoring are done (see Figure 2).
Table 5
ACADEMIC CLOUD SERVICES ADVANTAGES AND LIMITATIONS.
Advantages Limitations
Student • Free of cost • Resource Pooling and sharing • Back-Up and High Storage • Anywhere - anytime access • Green technologies • Speed and lack of Internet • Security and protection of sensitive data • Lack of knowledge to use the services
Teaching Staff • Deliver resource online • Academic reports • Availability • Communicating with students easily • Not all application can run on cloud • Services Downtime
Administrator • Data Management • Academic records • Admission process • Low IT staff • Less infrastructure cost • software updates. • Lack of confidence • Limited control • Vendor Lock-in
Researcher • Hardware utilization • Software availability • Huge information • Efficient testing • Pay-per-use • Reliability • Speed and lack of Internet • Security and protection of sensitive data
Developer • Appropriate Platform • Pay-Per-Use Work from anywhere • High Skills • High compute services • Not all application run on cloud • Organizational support and standards
In this paper, ESDB architecture is a scalable database for educational services. This architecture is designed using some main scalability techniques and methods which is:
• Provide horizontal scalability in different levels of the system to increase the system capacity by adding more and more resources to the existing system components. Also, the horizontal scalability provides the ability of using the old low performance equipment in a single high-performance system.
• The database is horizontally partitioned by rows using the Range - Hash combined method to get the
benefit of searching only one partition for the point queries and searching only a few partitions for the range queries.
• A cache memory attached to the database system to increase the response time of the already executed queries and stored procedures.
• Cluster server of Network Attached Server (NAS) that uses the master slave server architecture to horizontally scale up and down the system by adding or removing slave servers to the master server.
Application
Database Storage
Fig. 2. Cloud Computing Database Management Architecture.
Web Server, Application Server, Authentication and Customization
Cloud Database Storages
Backups, Data Encryption and Desk Monitoring
Table 5
ACADEMIC CLOUD COMPUTING SERVICES PROVIDERS.
Amazon Google Microsoft
Services Compute Elastic Compute Cloud (EC2) Database Amazon Relational Database Service (RDS), Dyna-moDB, SimpleDB, Elastic Cache. Google Apps Communication tools Gmail, Google Talk, Google Calendar Productivity tools Google docs, spreadsheets, presentations, iGoogle, Google Sites Database BigTable, BigQuery Microsoft Azure Microsoft Live@edu collaboration services, communication tools, mobile, desktop, and web-based applications • Office Live Workspace • Windows Live SkyDrive • Windows Live Spaces • Microsoft Outlook Live • Windows Live Messenger
Applications AWS Educate Google Scholar G Suite for Education Classroom, Gmail, Drive, Calendar, Vault, Docs, Sheets, Forms, Slides, Sites Azure in Education Microsoft Imagine ( .NET, Visual Studio, Office 365, Microsoft Graph), Visual Studio Dev Essentials (Virtual Machines, Storage, SQL Database), Microsoft Research, Machine Learning
Model IaaS PaaS PaaS
Limitations Create AWS account Problems (where most of the services needs a premium account) as students don't have a credit card even they have one, but they also worry about the charge from AWS Students may miss uses the credit card accidentally • Third-party apps that can integrate with the official Google apps and provide extra functionality can introduce new security risks into the normal use of G Suite. • Permissions can be confusing for many users and it's also common for files to be shared with users with more permissions than are necessary. • Azure is facing some technical problems according to the Microsoft Azure Status page. • Running and maintaining Azure drains and diverts valuable IT resources that are expensive and hard to find
Some Universities Uses the Services 2U, Baylor College of Medicine BCM, Carnegie Mellon University, Code.org, Coursera, Echo360, edX, Harvard University, Icahn School of Medicine at Mount Sinai, Notre Dame, UCAS, University of California at Berkeley, University of California, San Diego, University of Maryland, College Park, University of Oxford, University of San Francisco, University of Texas at Austin. University of Kufa, University Academy Keighley, Florida Atlantic University Lab Schools, Thurstable School, Sports College and Sixth Form Centre, Dainfern College, Parklands College & Christopher Robin Pre-Primary, Corlaer College, Leeds City College, Swallow Hill Community College, Uckfield Community Technology College, Thames Christian College, Barton Peveril Sixth Form College, College Community Schools, Meriden Public Schools. Duke University, Emory University, Thomas Jefferson University, University of Iowa, University of Washington, The University of Melbourne, Australian National University, Monash University, University of Wollon-gong.
In the main architecture of the ESDB, the database server clustered into a master slave server mechanism. The master server is a Network Attached Server (NAS) is a special purpose server that provides file-based data storage services over the network that can be configured remotely. This master server replicates all the data
updates over the slave servers. That ensures the distribution of the load over several servers where the user interacts with the slave servers, the slave servers respond to the read requests immediately and forward the write requests to the master NAS server (see Figure 3).
Fig. 3. . The main Educational Scalable Database ESDB architecture. The master server is supported with two extension software that can manage the partitioning and scalability progress. The first extension is the Interval
Partitioning Extension (IPE) that automatically creates new database partitions when the data exceeds the existing partition range that also can be configured. The second one is the Partition Advisor Extension (PAE), which recommends some partitioning and scalability strategy according to the workload.
At the educational level there are several kinds of information that can be processed in the cloud services, this information can vary from each other based on its uses and sensitivity. For that, this architecture provides three main levels of scalability, which are in the database, server, and cluster levels.
• Database scalability level: where the database can be horizontally partitioned and scaled by creating database partitions. This can be used when the data increases in a single unit of storage, for example when the student's records exceed the partition range it can be divided into new database partitions.
• Server scalability level: in this level of scalability, a new slave servers can be added or removed to scale out or scale in the system horizontally. This kind of scalability is recommended when a new instance of an existing set of data are required to be added for example, a new department or research center is introduced.
• Cluster scalability level: This can in general be implemented by duplicating the whole system architecture and replicating the data between them. This scalability mechanism is suitable for situations where a new instance of the system is required to be installed in another geographical region, for example, a new college, organization, or a different geographical region's campus.
7. CONCLUSIONS
The educational cloud computing services changed the general concept of the whole educational system. Where it became easy to access the resources from anywhere and anytime and reduces the cost regarding no software license is required. In this paper an educational cloud computing database model architecture named ESDB is introduced. This model is constructed around the horizontal scalability principle that ensures the ability of scaling the system by adding more resources to the existing system, besides that, the proposed scalable database architecture enables the using
of the old low performance equipment in a single highperformance system and as a result enables there will be no single unit of failure. An explanation and comparison of the main three cloud computing service providers (Google, Microsoft, and Amazon) with their advantages and limitations to conclude the suitable services to be applied for different requirements. The ESDB architecture horizontally scales the database by rows splitting using the Range - Hash combined method to get the benefit of searching only one partition for the point queries and searching only few partitions for the range queries. In addition to that, a cache memory attached to the database system to increase the response time of the already executed queries and stored procedures.
References
[1] Foster, I., Zhao, Y., Raicu, I., and Lu, S., Cloud computing and grid computing 360-degree compared, In Grid Computing Environments Workshop, 2008, Ieee, GCE'08 (pp. 1-10).
[2] Das, S., Scalable and elastic transactional data stores for cloud computing platforms, University of California, Santa Barbara.
[3] Abadi, D. J., Data management in the cloud: Limitations and opportunities. IEEE Data Eng. Bull., 32(1), 3-12.
[4] Gelogo, Y. E., & Lee, S., Database management system as a cloud service, International Journal of Future Generation Communication and Networking, 5(2), 71-76.
[5] Agrawal, D., El Abbadi, A., Das, S., & Elmore, A. J., Database scalability, elasticity, and autonomy in the cloud, In International Conference on Database Systems for Advanced Applications pp. 2-15, Springer, Berlin, Heidelberg, 2008.
[6] Agrawal, D., Das, S., & Abbadi, A. E., Data management in the cloud: challenges and opportunities, Synthesis Lectures on Data Management, vol. 4(6), pp. 1-138, University of California, Santa Barbara.
«етуушшим-шишаи» 2022 / technical science
39
[7] Narwal, M. S., Singh, M. S. P., Singh, M. L., & Kumar, M. A., Cloud Computing In Education., National Conference on E-Learning and E-Learning Technologies, 4th National Conference ELELTECH, India, Hyderabad.
[8] Pokorny, J., NoSQL databases: a step to database scalability in web environment. International Journal of Web Information Systems, vol. 9(1), pp. 69-82.
[9] Anandhi, R., & Chitra, K., Optimization Of Transaction Processing In Cloud Databases Through A Datacenter System Design Cracs To Improve Replica Consistency, Availability And Scalability, Journal of Theoretical & Applied Information Technology, vol. 62(3).
[10] Mathew, S., Implementation of cloud computing in education-A Revolution, International Journal
of Computer Theory and Engineering, vol. 4(3), pp. 473.
[11] Anandhi, R., & Chitra, K., A challenge in improving the consistency of transactions in cloud databases-scalability. International Journal of Computer Applications, vol. 52(2).
[12] Lakshminarayanan, R., Kumar, B., & Raju, M., Cloud computing benefits for educational institutions, arXiv preprint arXiv:1305.2616.
[13] Chandra, D. G., & Borah, M. D., Cost benefit analysis of cloud computing in education, In Computing, Communication and Applications (ICCCA) International Conference IEEE, pp. 1-6.
[14] Alattar M, Sali A., Strongly possible keys for SQL. Journal on Data Semantics, vol. 9(2), pp. 8599, 2020.
Ибрагимова Залина Майрбековна
Ассистент кафедры «Программирование и инфокоммуникационные технологии» Чеченского государственного университета им. А.А. Кадырова
Грозный, Россия Джамалдинова Марха Ахмадовна к.б.н., доцент кафедры технологии и дизайна Чеченского государственного
педагогического университета Грозный, Россия DOI: 10.24412/2520-6990-2022-31154-39-42 АНАЛИЗ БОЛЬШИХ ДАННЫХ УПРАВЛЕНИЯ ЦЕПОЧКАМИ ПОСТАВОК НА ОСНОВЕ IoT ПРИ УЧАСТИИ ПРОМЫШЛЕННЫХ ОТРАСЛЕЙ
Ibragimova Zalina Mayrbekovna
Assistant of the Department "Programming and Infocommunication Technologies" of the A.A. Kadyrov
Chechen State University Grozny, Russia Dzamaldinova Markha Akhmadovna Candidate of Biological Sciences, Associate Professor, Department of Technology and Design, Chechen
State Pedagogical University Grozny, Russia
BIG DATA ANALYSIS OF IoT-BASED SUPPLY CHAIN MANAGEMENT WITH THE PARTICIPATION OF INDUSTRIAL INDUSTRIES
Аннотация:
Цепочка поставок - это один из главных столпов производственных компаний, чья сообразительность может помочь бизнесу быть интеллектуальным. С этой целью использование инновационных технологий для придания ему интеллектуального характера всегда вызывает озабоченность. Интеллектуальная цепочка поставок использует инновационные инструменты для повышения качества, повышения производительности и облегчения процесса принятия решений. Интернет вещей (IoT) является одним из ключевых компонентов ИТ-инфраструктуры для развития интеллектуальных цепочек поставок, обладающих высоким потенциалом для обеспечения устойчивости в системах. Кроме того, Интернет вещей является одним из наиболее важных источников генерации больших данных. Большие данные и стратегии анализа данных как глубокое и мощное решение для оптимизации решений и повышения производительности быстро развиваются. По этой причине в данной статье предпринята попытка изучить информативные стратегии развития цепочки поставок, исследуя цепочку поставок в отраслях FMCG в качестве особого случая, и предоставить полную аналитическую основу для построения устойчивой интеллектуальной цепочки поставок с использованием больших данных на основе Интернета вещей.
Abstract:
The supply chain is one of the main pillars of manufacturing companies whose ingenuity can help businesses be intelligent. To this end, the use of innovative technologies to give it an intellectual character is always a concern. The intelligent supply chain uses innovative tools to improve quality, increase productivity and facilitate decision-making. The Internet of Things (IoT) is one of the key components of the IT infrastructure _ for the development of intelligent supply chains with high potential for ensuring sustainability in systems. In addition, the Internet of Things is one of the most important sources of big data generation. Big Data and data analysis strategies as a deep and powerful solution for optimizing solutions and improving performance are rapidly evolving.
