DISTRIBUTED SYSTEMS: CHARACTERISTICS, ARCHITECTURE, TYPES, OBJECTIVES, APPLICATIONS - TECHNOLOGY - 2025

The distributed systems are pieces of software used to coordinate the actions of multiple computers. This coordination is achieved through the exchange of messages, that is, with pieces of data that transmit information.

Distributed systems require concurrent components, a communication network, and a synchronization mechanism. They allow the sharing of resources, including software, by systems connected to a network. Therefore, the system is based on a network that connects the computers and handles the routing of messages.

a and b (distributed systems). b (parallel system). Miym / CC BY-SA (https://creativecommons.org/licenses/by-sa/3.0)

Distributed computing is an area of ​​computing that is responsible for analyzing distributed systems. Computer program that runs on a distributed system is called a distributed program.

In a context where there may be hundreds or thousands of computers, which is a common proportion in large Internet companies, it is very common that there are component failures, be it hardware, network, disks, etc., and the system must be prepared to face them at all times.

Data distribution

Distribution is key to being able to handle very large data conglomerates. It is necessary to achieve scalability, which is the means to maintain stable performance when data pools grow by adding new resources to the system.

On the other hand, distribution presents a series of technical problems that make the design and implementation of distributed computing and storage important to consider. One point to take into account is the risk of possible failures.

Characteristics of distributed systems

a and b (distributed systems). b (parallel system). Miym / CC BY-SA (https://creativecommons.org/licenses/by-sa/3.0)

Compatibility

The devices can work with different operating systems. This does not prevent them from always offering the same services to users. For this reason, all connected devices are compatible with each other.

Another fundamental issue is the design of the software, because it is also compatible with all the systems and users that are on each computer.

Fault tolerance

Being a single network with many computers, if any of its components fail, the others can continue to perform their function fully, avoiding errors quickly.

For this reason, distributed systems tend to provide a lot of confidence when working with them, because it is quite rare for the system to fail completely, because the tasks do not reside in a single device, but in different computers.

Middleware and API

Different processors use distribution middleware, helping to share different resources and capabilities to provide users with a coherent and integrated network. It also provides applications with a number of services, such as security and crash recovery.

Today, you hear more about application programming interfaces (APIs), which function as a gateway where applications can communicate. Applications do not need to know anything about other applications except their API.

Architecture

Client-server architecture

A particular computer called a server can perform certain tasks, which are called services. For example, serving files over the network, the ability to execute certain commands, or route data to a printer. The client is the computer requesting the services.

The computer known primarily for the service it provides may be called a print server, file server, and so on.

Peer-to-peer architecture

It assumes that each computer has similar capabilities and that no machine is dedicated to serving the others. An example of this is a set of microcomputers in a small office.

The network allows people to access each other's files and send emails, but no computer provides a specific set of services.

Multi-level architectures

For certain services it may make sense to have hierarchical connectivity. For example:

• A server when performing its tasks can contact another server of a different type
• The client-server architecture is a two-tier architecture.

Processor group architecture

An operating system could automatically start processes on idle computers and even migrate processes to systems with a higher number of available CPU cycles. In other cases, a user can manually start or move processes on available systems.

Types of distributed systems

Cluster computing

It is a collection of similar computers connected through a high-speed local area network. It is frequently used for parallel programming, where a single compute-intensive program runs in parallel on multiple computers.

Each cluster consists of a set of compute nodes that are monitored and managed by one or more nodes called masters.

Network computing

It is made up of nodes with marked differences in hardware and network technology. The current trend to have a specific configuration of nodes for certain tasks has led to greater diversity, which is the most common in network computing.

Cloud Computing

It is a set of virtualized resources housed in the data center of a cloud provider. Customers can establish a virtualized infrastructure to take advantage of a variety of cloud services.

To the user, it appears that he is renting his own exclusive computer. However, it is actually likely that you are sharing it with other customers. The same applies to virtual storage.

These virtualized resources can be dynamically configured, thus allowing scalability. If more computing resources are needed, the system can purchase more.

goals

Share resources

Be it storage facilities, data files, services or networks, you may want to share these resources between applications, for the sake of economy.

It is much cheaper to have a shared storage facility between multiple applications than to have to purchase and maintain storage for each one separately.

Abstraction

To hide that the processes and resources are distributed in several computers, possibly geographically dispersed. That is, the processes and resources are abstracted from the user.

Opening

It basically indicates that a distributed system is built with elements that can be easily integrated with other systems. Complying with standardized rules, any process with that interface can communicate with another process with the same interface.

Interoperability and portability

It refers to when two systems from different manufacturers can work together. Portability determines the extent to which an application made for system A can function on system B without modification.

Scalability

It is necessary when there is an increase in users who need more resources. A good example is the increase in the Netflix audience every Friday night.

It means dynamically adding more resources, such as increasing network capacity by allowing more video transmission and reducing it once consumption has normalized.

Applications

Erlang virtual machine

The LYME software package is based on Erlang and offers an alternative to LAMP. Shmuel Csaba Otto Traian / CC BY-SA (https://creativecommons.org/licenses/by-sa/3.0)

Erlang is a functional language that has great semantics for concurrency, distribution, and fault tolerance. An Erlang virtual machine handles the distribution of an Erlang application.

This model works by having many isolated processes, all with the ability to communicate with each other through a built-in messaging system.

The Erlang virtual machine can connect to other virtual machines that are in different places. This swarm of virtual machines runs a single application, handling the failures of one machine by scheduling it to run on another node.

Bittorrent

BitTorrent network operation. Mrjavi / CC BY-SA (https://creativecommons.org/licenses/by-sa/4.0)

It is one of the most widely used protocols for transferring large files on the web. The main idea is to facilitate the transfer of files between different peers on the network without having to go through a main server.

By using a BitTorrent client, you can connect to multiple computers around the world to download a file. A computer acting as a coordinator helps show the nodes on the network that have the desired file.

BitTorrent allows you to voluntarily host files and upload them for other users who want them. It is so popular because it was the first of its kind to offer incentives for contributing to the network.

Advantage

- More nodes can easily be added to the distributed system, that is, it can be scaled as needed.

- All nodes in the distributed system are connected to each other. Therefore, each of the nodes can easily share data with the other nodes.

- Resources such as printers can be shared with multiple nodes instead of being restricted to just one.

- The failure of one node does not lead to the failure of the entire distributed system. The other nodes can still communicate with each other.

Disadvantages

- Some messages and data may be lost on the network when moving from one node to another.

- It is difficult to provide adequate security in distributed systems because both nodes and connections must be secured.

- Network overload can occur if all nodes in the distributed system try to send data at the same time.

- The database connected to the distributed systems is quite complicated and difficult to manage compared to a single user system.

Examples of distributed systems

Distributed systems can be used in a large number of cases, such as electronic banking systems, massively multiplayer online games, and sensor networks.

StackPath

It uses a particularly large distributed system to power its content delivery network service. Each of its points of presence (PoP) has nodes, forming a worldwide distributed system.

StackPath stores the most recently and frequently requested content in the locations closest to the site that is being used.

By interconnecting virtual computers, in addition to taking advantage of the speed and agility of state-of-the-art computing, the system can handle thousands of simultaneous requests very quickly.

Internet

It is the largest distributed system in the world. Any user feels like a single system, even though it is made up of millions of computers.

Through the concept of abstraction, you have no idea where the data is stored, how many servers are involved, or how the information gets to the browser. The browser dissolves the complexity of the Internet.

This also applies to applications such as Gmail email or any other application that can be used. Every person interacts with distributed applications on a daily basis.

References

1. Paul Krzyzanowski (2018). Distributed Systems. Taken from: cs.rutgers.edu.
2. Catherine Paganini (2019). Primer: Distributed Systems and Cloud Native Computing. The New Stack. Taken from: thenewstack.io.
3. International University of Valencia (2020). Distributed systems, characteristics and classification. Taken from: universidadviu.com.
4. David Meador (2018). Distributed Systems. Tutorials Point. Taken from: tutorialspoint.com.
5. Robert Gibb (2019). What is a Distributed System? Taken from: blog.stackpath.com.
6. Stanislav Kozlovski (2018). A Thorough Introduction to Distributed Systems. Free Code Camp. Taken from: freecodecamp.org.