Download to read offline
More Related Content
Similar to Chapter 3 Naming in distributed system.pptx
Chapter 3 Naming in distributed system.pptx
- 1. Process and Threads Naming in DistributedSystems Chapter 3
- 2. Outline • Process andThreads in DS • Client and Server • Naming Basics • Name Resolution • Flat Naming vs. Hierarchical Naming: • Name Services • Name Spaces
- 3. Threads and Processin Distributed Systems • In distributed systems, processes and threads are fundamental concepts that play a crucial role in enabling concurrent execution, efficient resource management, and overall system performance. • They provide the means to structure and manage the execution of tasks and programs in a distributed environment. • Here's an introduction to processes and threads in distributed systems:
- 4. Cont’d… • Processes: • Aprocess is an independent and self-contained program execution unit within an operating system. • Each process operates in its own isolated memory space and has its own system resources, making it independent of other processes. • Key characteristics of processes in distributed systems include: • Isolation: Processes are isolated from one another, meaning that the failure or misbehavior of one process does not affect others. This isolation enhances system security and stability.
- 5. Cont’d… • Resource Management:Each process has its own dedicated resources, such as CPU time, memory, and file handles. This allows for efficient resource management and prevents resource contention. • Security: Processes can be individually protected and managed. Access control mechanisms can be applied to restrict unauthorized access and interactions between processes. • Parallelism: Processes enable parallel execution on multi-core processors, which is essential for leveraging the full potential of modern hardware.
- 6. Cont’d… • Processes aretypically created using system calls provided by the operating system. In distributed systems, processes can communicate with each other through inter-process communication (IPC) mechanisms like pipes, sockets, shared memory, or message queues. • Threads: • A thread is a lightweight unit of execution within a process. • Threads share the same memory space and resources within a process and are sometimes referred to as "lightweight processes." Key characteristics of threads in distributed systems include:
- 7. Cont’d… • Concurrency: Threadsenable concurrent execution of multiple tasks within a single process, taking advantage of multi-core processors. This concurrency improves system responsiveness and throughput. • Efficiency: Threads have lower overhead compared to processes because they share resources and memory. This makes them more efficient for tasks that require frequent context switching or parallelism. • Shared State: Threads within the same process can easily share data and communicate with each other, making them suitable for collaborative and closely related tasks.
- 8. Cont’d… • In distributedsystems, processes and threads are essential for various reasons, including: • Handling concurrent requests: Processes and threads can handle multiple incoming requests simultaneously, improving system responsiveness and throughput. • Resource utilization: By distributing tasks across processes or threads, distributed systems can utilize system resources efficiently, reducing idle time. • Isolation and security: Processes provide strong isolation, while threads can be useful for sharing data and resources within a controlled context. • Parallelism and scalability: Processes and threads enable distributed systems to take advantage of multi-core processors, improving performance and scalability.
- 9. Clients and servers •Clients and servers in distributed systems use processes and threads to handle various tasks and provide services efficiently. • Here's how clients and servers employ processes and threads to achieve their respective roles: • Servers: • Process Per Connection Model: • Servers often use a "process per connection" model, where a new process is created for each incoming client connection. • This model provides strong isolation, ensuring that client interactions do not interfere with each other. • Each process can be dedicated to handling a single client request or session, allowing the server to maintain multiple concurrent connections.
- 10. Cont’d… • Thread PerConnection Model: • An alternative to the process per connection model is the "thread per connection" model. • In this approach, a new thread is created for each incoming client connection within a server process. • Threads share the same process address space but have their own execution context. • This model is more lightweight than creating separate processes and is suitable for handling a large number of simultaneous connections.
- 11. Cont’d… • Worker Threadsor Processes: • Servers often employ a pool of worker processes or threads to handle incoming client requests. When a new client connection is established, the server assigns the task to an available worker thread or process. This approach balances the load and efficiently manages resources. • Task Parallelism: • Within a server, multiple threads or processes may be responsible for different tasks, such as listening for incoming connections, processing client requests, and handling database queries. This division of labor improves the server's overall performance and responsiveness.
- 12. Cont’d… • Shared StateManagement: • When multiple threads or processes are used within a server, mechanisms for managing shared state and data are critical. • This may involve using synchronization techniques, such as locks or semaphores, to prevent data corruption and ensure thread-safe access to resources.
- 13. Client • Concurrent Requests: •Clients may use threads to issue multiple concurrent requests to remote servers. • Each thread can send a request, receive a response, and manage its own session with the server. • This concurrency allows clients to efficiently use the network and reduce latency. • Asynchronous Operations: • In some cases, clients use asynchronous programming models, such as asynchronous I/O or event-driven architectures, to issue requests to servers. These models allow clients to continue executing other tasks while waiting for responses from the server, improving overall client responsiveness.
- 14. Cont’d… • Parallelism inDistributed Computing: • In distributed computing scenarios, clients may employ threads to parallelize data processing tasks or computation across multiple servers. • This parallelism accelerates data analysis, simulations, or other computational workloads. • Load Balancing: • Clients can use multiple threads or processes to interact with multiple server instances or replicas, distributing the load and ensuring redundancy. • Load balancing mechanisms help optimize resource utilization and enhance system reliability.
- 15. Cont’d… • Resource Management: •Clients may create and manage threads to handle various tasks, such as user interfaces, data processing, and communication with servers. • Effective resource management is essential to ensure that client applications remain responsive and efficient. • Clients and servers collaborate to provide distributed services effectively by using processes and threads for concurrent execution, load distribution, and resource management.
- 16. NEXT… Naming in DistributedSystems
- 17. Introduction to Naming •In distributed systems, naming is the process of assigning names or labels to various entities, such as servers, resources, processes, and services. • Names are used to uniquely identify and reference these entities within the system. • In distributed systems, various entities are named to uniquely identify and reference them within the system. • Naming entities serve as a fundamental component in facilitating communication, resource location, and coordination within distributed environments.
- 18. Cont’d… • Here arecommon naming entities in distributed systems: • Hosts and Servers: • Hosts and servers in a network are often given names to identify them. For example, a web server may be named "webserver1." • Services: • Services or software components may be given names to make them accessible and to locate them. • This is common in microservices architectures, where each service has a unique name for discovery and communication. • Processes and Threads: • Processes or threads running on a host can be named for identification and coordination. • This is particularly important in distributed computing when coordinating tasks among different processes.
- 19. Cont’d… • Distributed Objects: •In distributed object-oriented systems, objects are named to enable remote method invocation. Each object has a unique name or reference that clients can use to interact with it. • Resources: • Various types of resources, such as printers, storage devices, or sensors, are often named in a distributed environment. Naming allows users or applications to discover and access these resources. • Data and Databases: • Databases and the data within them can be named for querying and retrieval. Names are used to identify tables, records, or specific data items.
- 20. Name Resolution: • Nameresolution is the process of mapping a name to the corresponding network address or location. • When components in a distributed system need to communicate with each other, they use names to initiate connections. • Name resolution is essential to determine where the named entity can be found in the network. • Name resolution is the process of mapping a human-readable name, such as a hostname or service name, to a network address or location that can be used for communication in a distributed system.
- 21. Cont’d… • Name resolutionis an essential part of networking and distributed systems, and it ensures that entities can locate and communicate with each other by using names instead of numerical addresses. • Here's how name resolution works in typical distributed systems: • Name Request: • When a component in a distributed system wants to communicate with another component, it uses a human-readable name (e.g., a domain name, a service name, or a hostname) to identify the target entity.
- 22. Cont’d… • Local CacheLookup: • To expedite the name resolution process, the requesting component may first check its local cache, which stores previously resolved name- to-address mappings. • If the name is found in the cache and is still valid, the component can use the cached address for communication, avoiding the need for external resolution.
- 23. Cont’d… • Local Configuration: •If the name resolution request is not found in the local cache, the component may check its local configuration files or settings. Some systems allow administrators to configure custom mappings of names to addresses locally. • Name Service Query: • If the name resolution request cannot be resolved locally, the component contacts a name service, which is responsible for maintaining a central or distributed database of name-to-address mappings. • Common examples of name services include DNS (Domain Name System) for internet domain names or service discovery systems in microservices architectures.
- 24. Cont’d… • Caching inName Services: • Name services often have their own caching mechanisms to store resolved name-to-address mappings temporarily. This helps reduce the load on the name service and speeds up future name resolution requests for the same names. • Iterative Resolution (DNS): • In the case of DNS, an iterative query involves multiple DNS servers. The client first contacts a local DNS resolver, which may not have the answer but can refer the client to authoritative DNS servers. The client follows the referrals, querying various DNS servers until it reaches the authoritative server for the domain in question.
- 25. Name Services: • Nameservices are components or systems responsible for maintaining a mapping of names to network addresses. • These services act as a central repository for name-to-address mappings. • For example, the Domain Name System (DNS) serves as a global name service for the internet, mapping domain names to IP addresses. • Various systems and protocols are used for implementing name services in distributed environments. • These systems are responsible for managing the mapping of human- readable names to network addresses or locations.
- 26. Cont’d… • Here aresome of the common systems and protocols used for name services: • Domain Name System (DNS): • DNS is one of the most well-known and widely used name services. It's the naming system for the internet, translating domain names (e.g., www.example.com) into IP addresses. • DNS is hierarchical, with a distributed architecture that includes authoritative name servers, caching resolvers, and recursive resolvers. • Lightweight Directory Access Protocol (LDAP): • LDAP is a directory service protocol that provides name services for directory- based systems. It's commonly used in enterprise environments for storing and retrieving information about users, devices, and resources.
- 27. Cont’d… • Service DiscoverySystems • Service discovery systems are designed for microservices architectures, allowing services to register and discover one another. • They maintain a database of service names and their locations, providing load balancing and failover capabilities. Examples include Consul, etcd, and Apache ZooKeeper. • Distributed Hash Tables (DHTs): • DHTs are data structures used for decentralized and distributed name services. They allow distributed entities to map keys (names) to values (data) in a decentralized manner. DHTs are commonly used in peer-to-peer systems, and Chord is an example of a DHT-based name service.
- 28. Name Spaces: • Distributedsystems often define separate name spaces for different types of entities. • For example, in a distributed file system, there may be distinct name spaces for files and directories. • This separation helps manage and access different types of entities. • A namespace, in the context of distributed systems and computing, is a logical or virtual container that provides a way to organize and manage the naming of entities, resources, or identifiers within a system
- 29. Cont’d… • Namespaces aredesigned to prevent naming conflicts and improve organization by partitioning the global name space into smaller, more manageable units. Here's how namespaces work: • Organizing Entities: • Namespaces are used to organize and group entities, resources, or identifiers. Each namespace represents a distinct, isolated container for naming. This organization is especially useful in large and complex distributed systems where a variety of entities need to be uniquely identified. • Hierarchical Structure: • Namespaces are often structured hierarchically, forming a tree-like or parent-child relationship among namespaces. This hierarchy allows for the nesting of namespaces within one another, creating a logical structure.
- 30. Cont’d… • Namespace Resolution: •When a distributed system receives a request or reference that includes a name, the system uses the concept of namespaces to determine which namespace the name belongs to. Namespace resolution helps identify the appropriate container for the name, allowing the system to access the entity within the correct context. • Eliminating Naming Conflicts: • Namespaces effectively eliminate naming conflicts by providing a context or scope for each set of names. Entities with identical names in different namespaces do not conflict with each other, as they are considered distinct within their respective containers.
- 31. Cont’d… • Domain NameSystem (DNS): • The DNS is a classic example of a namespace. In the DNS, domain names are organized hierarchically, with top-level domains like .com, .org, and .net, and subdomains like example.com or subdomain.example.com. Each domain or subdomain represents a namespace. For example: • "example.com" and "another.com" can coexist with the same subdomain name, such as "www," without conflicts, as they belong to different namespaces.
- 32. Cont’d… • File Systems: •File systems use namespaces to organize files and directories. In this case, each directory is a namespace, and file names must be unique within the context of their parent directory. For example: • In a file system, you can have two directories with files named "document.txt." Each directory serves as a separate namespace, ensuring that "document.txt" is unique within its parent directory
- 33. Types of Naming •In distributed systems, various types of naming schemes and approaches are used to uniquely identify and address entities, resources, or services. • These naming methods are designed to meet the specific needs of different distributed systems. • Flat Naming: • Explanation: Flat naming, as the name suggests, uses a single, global name space where all entities have unique names. There are no hierarchical or structured elements in the names. Each entity is identified by a distinct name. • Use Cases: Flat naming is suitable when the number of entities is limited and the naming scheme is simple. It is often used in small-scale, non- hierarchical systems.
- 34. Hierarchical(Structured) Naming: • Explanation:Hierarchical naming organizes names in a tree-like structure, where each name consists of components that form a path from the root to the entity. • This hierarchy allows for categorization and organization of entities, making it easier to manage and locate them. • Use Cases: Hierarchical naming is useful in large-scale distributed systems where entities can be grouped and organized logically. • Examples include the Domain Name System (DNS) for internet domains or file systems with directories and files.
- 35.
- 36. Semantic(Attribute-Based) Naming: • Explanation:Semantic naming uses names that convey meaning or attributes about the entity or resource they represent. • Names are chosen based on their semantic relevance, making them more human-readable and self-explanatory. • This approach aims to make names self-descriptive and meaningful, which can be more user-friendly and intuitive. • Use Cases: Semantic naming is often used in contexts where the naming scheme needs to be user-friendly or self-documenting. • For example, in a document management system, files may be named based on their content or purpose.
- 37. Cont’d… • Example: Ina distributed system for managing scientific papers, attribute-based naming may involve naming papers based on their attributes, such as the title, author, and publication year. For example, a paper might be named "Astronomy_Guide_to_Stars_2023.pdf." In this name: • "Astronomy" represents the subject or category of the paper. • "Guide_to_Stars" is the title of the paper. • "2023" is the publication year.
- 38. The End!!! Thankyou foryour Attention!!!