Computer Science Interview Questions and Answers 2026

Let us see first what is a kernel: - 

Kernel: The core of an operating system is its kernel (OS). It serves as a link between a computer's hardware and its software programmes. The system's resources, including the CPU, memory, and input/output devices, are managed by the kernel.

Now, let’s see various types of kernels in operating system. 

• Monolithic kernel
  • The kernel itself contains all functions.
  • It is large and bulky.
  • A lot of memory is needed to operate.
  • Less reliable, if one module fails, the entire kernel goes down.
  • High performance due to quick communication.

Example: Linux, Unix, and MS-DOS.

• Micro Kernel
  • The kernel only contains major functions.
  • Memory management
  • Process management
  • File management and IO management are in user space.
  • It is smaller in size.
  • It is more secure. 
  • Performance is comparatively slow.
  • Inefficient transition between kernel mode and user mode.

Example: L4 Linux, Symbian OS, MINIX, etc.

• Hybrid Kernel
  • In this, file management in user space and rest in kernel space.
  • It contains monolithic speed and design.
  • It contains micro's modularity and stability.
  • IPC occurs as well, although with lower overheads. 

Example: MacOS, Windows NT/7/10 

Apart from that there are more types of kernels that is comparatively less used in OS.

• Nano Kernel
• Exo Kernel

Thrashing occurs when a computer system has insufficient memory and spends a lot of time switching data between the disk and memory rather than performing tasks.  

The process will immediately page-fault if it does not have the necessary number of frames to support the active pages. It must now swap out a page at this point. It must, however, replace a page that will be immediately required again because all of its pages are currently being used. As a result, it faults frequently, replacing pages that it needs to bring back in right away.  Thrashing is the name for this high-paging action.  

Thrash is the term used in computer science to describe the poor performance of a virtual memory (or paging) system caused by the repetitive loading of the same pages when there is not enough main memory to retain them in mind.

The Causes of Thrashing are as follows:  

• High level of multiprogramming. 
• Lack of frames.  

Recovery from thrashing:   

1. Use Don't Allow Bit: The "Don't Allow Bit" can be used to restrict how much memory each process or application is allowed to consume in order to avoid thrashing. The "Don't Allow Bit" can be set to prevent additional memory allocation when a process or application reaches its memory limit and forces the process to release any unused memory. 
2. Adding more RAM: The amount of thrashing can be decreased by increasing the system's random access memory (RAM), which gives it more memory to work with. 
3. Increase the size of the swap space: Increasing the swap area's size can help the system experience less thrashing if it currently has a low swap space. 
4. Reduce the number of processes: Reduce the number of programmes operating at once or shut down any unnecessary processes to lessen thrashing. 
5. Upgrade the hardware: The amount of time of thrashing can be decreased by installing a solid-state drive (SSD) or upgrading to a faster hard disc. 

Banker’s algorithm is used to avoid deadlock. 

The term was chosen because the algorithm might be implemented in a banking system to make sure that the bank never distributes its funds in a way that prevents it from meeting the needs of every customer. There should be a maximum number of each resource type's instances.  

The maximum amount that may be allocated safely to determine the allocation for all resources is tested using this algorithm. Before deciding whether or not to continue allocating, it also checks for all potential activities.  

For instance, there are G dollars total in the accounts of X number of account holders at a particular bank.  

A car loan is processed by the bank using software subtracts the loan amount for a car purchase from the total amount of money the bank possesses (G+ Fixed Deposit + Monthly Income Scheme + Gold, etc.).  

Additionally, it verifies whether the difference exceeds G or not. Even if every account holder withdraws funds at once, the auto loan won't be processed until the bank has enough cash.  

Friend class is a class that has been designated as a friend can access the private and protected members of other classes. Sometimes it is advantageous to provide a certain class access to another class's private members.  

A friend class can access both private and protected members of the class in which it has been declared as a friend.  

Friend Functions:  

Friend Functions, like friend class, can also access private and protected members.  

In C++, a friend function is a special function that, although not belonging to a class, has the ability to access its private and protected data.  

• A friend function is a non-member function of a class, which is declared by the keyword "friend" so that it can access all the members of that class. 
• The friend function's function declaration, not the definition, is the only place the term "friend" is used.  
• Neither the object's name nor the dot operator is utilised when the friend function is invoked. The object, whose value it wants to access, may be accepted as an argument.  
• Any portion of the class, such as the public, private, or protected, can be used to declare a friend function.  

The following are some crucial details about friend functions and classes:  

• Friends should only be utilised in specific circumstances. The utility of encapsulating different classes in object-oriented programming is diminished if too many functions or external classes are specified as friends of a class with protected or private data.  
• Friendships are not reciprocal. Class B does not immediately become a buddy of class A just because class A is.  
• Friendship is not inherited.  
• The concept of friend classes is not supported by Java.  

Merits of using Friend Class/Function:  

• Friend functions can access members without having to inherit the class.  
• The friend function connects two classes by having access to their personal information.  
• It can be declared in the class's protected, private, or public part.  

Demerits of using Friend Class/Function:  

• A class' private members can be accessed by friend functions from outside the class, which is against the law of data hiding.  
• Friend functions cannot do any run time polymorphism in its members. 

Abstraction: Abstraction is showing only the essential information and hiding all the other unnecessary things and details.

Example: 

Let’s say a man is driving a vehicle. The man only knows that pressing the accelerators will make a car go faster or that applying the brakes will cause the car to stop, but he has no idea how pressing the accelerator actually increases speed. He also has no idea how the accelerator, brakes, and other car controls are used inside the vehicle. 

Abstracts Class:

An abstract class is declared by the keyword “abstract”. It can contain abstract as well as non-abstract methods. An abstract class is one that can't be created on its own. It offers a base class definition from which subsequent classes can inherit and serves as a blueprint for developing concrete subclasses. 

Creation: abstract type name(parameter-list);

Properties of Abstract Class:  

1. All classes must be declared abstract if they have one or more abstract methods.
2. There cannot be any abstract class objects.
3. We cannot declare abstract methods or abstract static constructors.
4. We can declare static methods in abstract class because there can be no objects for abstract class. 
5. Any subclass of an abstract class must either be declared abstract or implement every abstract method in the superclass.
6. Abstract classes may have as much implementation as they see appropriate, including concrete methods (methods with bodies).

A must-know for anyone heading into a CS interview, this question is frequently asked in CS interview questions. Http and Https are protocols rather they are application-layer protocols used for transmitting data over the Internet.

Let us see what is a protocol and application-layer protocol. 

• Protocol: A protocol is a set of rules for structuring and processing data in networking.
• Application-layer Protocol: A protocol that works at the application level of the network communication model is known as an application-layer protocol. Some examples apart from HTTP and HTTPS are FTP (File Transfer Protocol), SMTP (Simple Mail Transfer Protocol), DNS (Domain Name System) and SSH (Secure Shell). 
• HTTP: Files like HTML pages and photos are requested and sent from a web server to a client via HTTP. It is a stateless protocol, which implies that between requests, it keeps no track of the status of a connection. For networked, collaborative, hypermedia information systems, HTTP is utilized as an application-layer protocol. The HTTP protocol is used to interact between website visitors and server farms. World Wide Web uses it to control communications between web servers and browsers.
• HTTPS: SSL/TLS encryption is used in HTTPS, a secure variant of HTTP, to safeguard the data being communicated between a client and a server. As a result, sending sensitive data over the Internet, including login passwords or financial information, via HTTPS is more secure. 

If a particular HTTP request has indeed been correctly answered, it will be indicated by a response status code.

Five categories have been used to categorize responses:

• Informational answers (100 – 199)
• Successful reactions (200 – 299)
• Messages of redirection (300 – 399)
• Client response errors (400 – 499)
• Server error (500 – 599)

IP address: An IP address (Internet Protocol address) is a unique numerical identifier given to each connected device to a computer network that makes use of the Internet Protocol. A host's or network interface's identification and position inside the network are the two primary purposes of an IP address. Although IP addresses are often expressed as a string of alphanumeric letters, computers interpret them as binary codes (1s and 0s).

There are two main versions of IP addresses: IPv4 and IPv6.

IPv4:

• 32-bit addresses are supported by IPv4, however, end-to-end connection integrity is not possible with IPv4 due to the inability of manual and Dynamic Host Control Protocol (DHCP) address setup.
• 4.29 109 address space can be generated. Application-dependent security feature
• IPv4 addresses are represented using decimal numbers.
• Senders and forwarding routers execute fragmentation
• IPv4 does not have an encryption or authentication capability. 
• The four fields that makeup IPv4 are separated by dots (.)
• There are five distinct classifications for IP addresses in IPv4. Classes A, B, C, D, and E are available.

IPv6:

• IPv6 addresses are 128 bits long and offer auto- and renumbering address setup.
• Connection integrity is achievable with IPv6 end-to-end.
• The IPv6 protocol contains built-in security features like Internet Protocol Security (IPSEC). Its address encoding is in hexadecimal. It has a huge address space that can yield 3.41038 address space. It also supports multicast and multicast data delivery schemes. It also has encryption and authentication. It has a set header size of 40 bytes.
• 8 fields make up IPv6, each field is separated by a colon (:)
• IP addresses in IPv6 do not fall into any categories.
• Variable Length Subnet Mask (VLSM) is incompatible with IPv6.

They all are various types of networking device. 

1. Repeater: At the physical layer, a repeater functions. Its task is to regenerate the signal over the same network before it weakens or becomes corrupted in order to increase the amount of time that it may be sent over the same network. Repeaters do not enhance the signal, which is a crucial distinction to make. When the signal deteriorates, they replicate it piece by piece and restore it to its initial intensity. The gadget has two ports.
2. Hub: Essentially, a hub is a multiport repeater. A hub joins several wires that come from several branches, like the connector in a star topology that joins various stations. Data packets are delivered to all connected devices since hubs are unable to filter data. In other words, all hosts linked by Hub continue to share a single collision domain. Additionally, they lack the intelligence to choose the optimum route for data packets, which results in waste and inefficiency.
3. Bridge: The data link layer is where a bridge functions. A bridge is a repeater with the additional capability of content filtering via source and destination MAC address reading. Additionally, it is utilised to link two LANs that use the same protocol. It is a 2-port device since it has a single input and output port.
4. Switch: A switch is a multiport bridge with a buffer and a design that can increase its performance and efficiency (more ports mean less traffic). A data link layer device is a switch. The switch may carry out error checking before forwarding data, which makes it incredibly efficient because it only forwards good packets to the right port and does not transmit packets with mistakes. To put it another way, the switch separates the hosts' collision domains while keeping the broadcast domain constant.
5. Routers: Similar to a switch, a router directs data packets depending on their IP addresses. Mainly a Network Layer device, the router. LAN and WAN connections are often made via routers, which also decide how to route data packets based on a dynamically updated routing table. The broadcast domains of hosts linked by a router are divided.
6. Gateway: A gateway, as its name indicates, is a passageway that connects two networks—which may use various networking models—together. They essentially serve as the messengers who interpret and transport data from one system to another. Protocol converters are another name for gateways, which may work at any network layer. In general, gateways are more complicated than switches or routers.

1. Socket: A socket is one endpoint of a two-way communication channel between two networked programmes. By creating named contact points between which the communication takes place, the socket mechanism offers a form of inter-process communication (IPC).
2. Subnet: A network inside another network is known as a subnet, or subnetwork. Networks become more effective with subnets. Network communication can go a shorter distance to its destination without using extraneous routers thanks to subnetting. Networks operate more effectively when communications move as directly as possible, much like the postal service. In order to avoid sending data packets on an inefficient path to their destination, a network sorts and routes data packets it receives from another network according to subnet.
3. Subnet Mask: Similar to an IP address, a subnet mask is only used internally within a network. Subnet masks are used by routers to direct data packets to the correct location. Data packets travelling over the Internet do not contain subnet mask information, instead, they simply contain the destination IP address, which a router will match with a subnet.
4. Protocol: A protocol is a set of rules for structuring and processing data in networking.
5. Cookie: A cookie is a piece of information from a website that is saved in a web browser for subsequent retrieval by the website. Cookies are used to let a server know whether visitors have visited a specific website again. A cookie gives information and enables the website to display customised settings and tailored content when visitors return. Additionally, cookies keep track of user preferences, the contents of the shopping cart, and login or registration information. This is done so that any data supplied in a previous session, or any specified preferences may be simply retrieved when visitors revisit websites.
6. Star Topology: A network structure known as a "star topology" has all nodes connected to the hub, which serves as the centre node, through cables. The hub might be either active or passive. Repeaters are found in active hubs, but passive hubs are regarded as non-intelligent nodes. To the central node, which serves as a repeater for data transmission, each node has a reserved link.
7. Mesh Topology: Mesh topology is a type of topology where every node is linked to every other node by a network channel. There is a point-to-point link in mesh topology. It can link n nodes using n(n-1)/2 network channels. Routing and flooding are two of the data transfer methods available in mesh topology. The nodes in the routing method each have a routing logic, such as the logic for the quickest path to the target node or the logic to prevent routes with broken connections. The network nodes in the flooding strategy all get the same data. We no longer require routing logic as a result. Although this method strengthens the network, it adds unneeded load.
8. Tree Topology: A tree topology is a topology in which all nodes are connected to the topmost node, or root node, in a hierarchical manner. It is also known as hierarchical topology for this reason. Three degrees of hierarchy are included in tree topology. Wide Area Network uses tree topology. It is a fusion of both Bus and Star topologies.
9. Hybrid Topology: A network topology that combines two or more distinct topologies is known as a hybrid topology. It is an expensive topology but one that is also dependable and scalable. It benefits and suffers from the topologies that were utilised to construct it.  

• Static Memory Allocation:
  • Allocation is carried out using Stack Memory. 
  • The variables' memory space is reserved permanently until the programme or function call is finished.  
  • During the compilation step, memory allocation takes place. 
  • Declared memory is retained during the entirety of the program's execution.   
  • When compared with dynamic memory allocation, it is simple.  
  • In this allocation, allotted memory cannot be utilised again.  
  • It has ineffective memory management. 
  • In static memory, the memory size cannot be modified after it has been allocated to a programme. 
• Dynamic Memory Allocation: 
  • During the runtime step, memory allocation takes place.  
  • The new keyword only uses memory allocation for the variables when it is necessary.  
  • Heap memory is used for allocation.  
  • If memory is allocated for a programme in dynamic memory, the memory size can be altered later. 
  • It has improved memory management.  
  • If the memory that was allocated is no longer needed, it can be utilised again.  
  • When it comes to defining multi-dimensional arrays, it is more difficult.  
  • Memory that has been declared can be released, reused, and assigned to other purposes.

A theoretical framework known as the OSI (Open Systems Interconnection) model specifies and standardises a communication system's operations in seven different levels. 

1. Physical layer: It addresses with the functional, electrical, and mechanical sides of the physical connection between devices, including the bit-transmission through a physical media like copper wire or optical fibre.
2. Data Link Layer: It offers flow management, medium access control, and error detection and correction, providing dependable transmission of data frames through the physical layer.
3. Network Layer: For reliable data frame transmission via the physical layer, network layer offers error detection and correction, flow control, and medium access control.
4. Transport Layer: Data transmission from the source host to the destination host must be error-free and end-to-end. This is the responsibility of the transport layer.

The functions of the transport layer are: 

It makes it possible for the hosts who are communicating to have a discussion. 

It can also provide flow control, verification, and error checking. 

1. Session Layer: The Open System Interconnection (OSI) model's fifth tier is the session layer. This layer enables users to create active communications sessions with one another across many devices. Establishing, maintaining, synchronising, and terminating sessions between end-user applications are its responsibilities. In the session layer, streams of data are received, further tagged, and then correctly resynchronized. This prevents message ends from being originally severed and additional data loss. In a nutshell, this layer creates a link between the session components. 
2. Presentation Layer: The Open System Interconnection (OSI) model's sixth tier is called the Presentation Layer. Because it acts as a data translator for the network, this layer is also known as the translation layer. In order to transfer the data across the network, this layer extracts and modifies the data that it gets from the Application Layer. This layer's primary duty is to offer or specify the data format and encryption. The presentation layer is sometimes known as the syntax layer because it is in charge of ensuring that the data it receives or sends to other layers has the correct syntax.
3. Application Layer: It is in charge of determining the structure and meaning of data that is sent between apps as well as providing services to the user. 

Note: In TCP/IP model, its application layer combines the features of the OSI model's session layer, presentation layer, and application layer. 

The idea of concurrency control falls within the database management system's Transaction category (DBMS). A DBMS procedure lets us manage two concurrent processes so that they may run without conflicting with one another, which happens in multi-user systems.

Executing many transactions simultaneously is the simplest definition of concurrency. To improve time efficiency, it is necessary. Inconsistency develops when many transactions attempt to access the same piece of data. Data consistency requires the use of concurrency control.

If we use ATMs as an example, numerous people cannot withdraw money simultaneously at various locations if concurrency is not used. We require concurrency here.

The following are concurrency control's advantages: 

• There will be less waiting. 
• The turnaround time will shorten. 
• The use of resources will increase. 
• Improved system performance & efficiency. 

Main problems in using Concurrency: Updates will be lost: Updates won't be saved if one transaction makes modifications and another one undoes them. The changes of one transaction override those of another. Uncommitted Dependency or Dirty Read Problem: When a variable is updated in one transaction and its value is deleted in a subsequent transaction, the variable that was updated in the first transaction is not updated or committed, giving us false values or the variable's previous value. Inconsistent retrievals: When one transaction updates several distinct variables and another transaction is in the process of updating those variables, the problem is that the same variable appears inconsistently in several contexts. 

Concurrency control strategies: The strategy for concurrency control are as follows:
 
Locking: Lock ensures that a current transaction has sole access to certain data objects. It initially obtains a lock to get access to the data items, and once the transaction is over, it releases the lock.
 
The various lock types are as follows:

• Shared Lock [Transaction can only read the values of data items]. 
• Exclusive Lock [Used for read-only and write-accessible data item values] 

Relationship constraints are rules that specify how different database entities relate to one another. In order to ensure that data is saved and retrieved consistently and properly, they are used to enforce business rules and preserve data integrity. 

There are two main types of binary relationship constraints. 

1. Mapping Cardinality 
2. Participation Constraints 

Mapping Cardinality / Cardinality Ratio:
 
Number of entities that may be connected to another entity through a connection is known as the mapping cardinality or cardinality ratio.

Types of Mapping Cardinality:

One to One: An entity in entity set A associates with no more than one entity in entity set B. And a B entity can only be related to a single A entity.

E.g., Let’s take an example of student database, where each student has only one roll number.

In this example, the relationship between the "Student" entity and the "Roll Number" entity would be a one-to-one relationship, as each student has only one roll number and each roll number is assigned to only one student. This can be represented using the following mapping cardinality: 

• One student has one roll number (1:1) 
• One roll number is assigned to one student (1:1) 

One to Many: N entities in B are connected to an entity in A. While only one of the entities in A is connected to the entity in B.

E.g., Let's take the example of a bookstore database. Each book (entity B) can be written by only one author (entity A), but an author can have written multiple books. Therefore, the relationship between authors and books would be a one-to-many relationship. 

This can be represented using the following mapping cardinality: One author can write many books (1:N) 

Many to one: Entity in A is connected to at most one entity in B. However, N entity in A can be connected to entity in B.
 
E.g., Let’s take the same example of student database, where many students can have the same academic advisor. 

In the case of the student database, many students could potentially have the same academic advisor (entity B), but one academic advisor can advise many students (entities A). Therefore, the relationship between students and academic advisors would be a many-to-one (N:1) relationship.

Many to Many: N entities in B are connected to an entity in A. While N entity in A is also connected to entity in B.

E.g., Again, let’s take an example of student database, where many students can be taught by many teachers. 

In the case of the student database where many students can be taught by many teachers, each student can be associated with multiple teachers (entities B), and each teacher can teach multiple students (entities A). Therefore, the mapping cardinality for this relationship would be many-to-many (N:N). 

Participation Constraints: Participation constraints can be defined as the minimum and maximum participation of an entity in a relationship. They determine whether an entity must participate in a relationship or not. 

There are 2 types of participation constraints: - 

• Total Participation 
• Partial Participation 

• Firewall: A form of cybersecurity equipment called a firewall is used to control network traffic. The usage of firewalls may be employed to isolate network nodes from internal and external traffic as well as from apps specifically designed. Firewalls can be hardware, software, or cloud-based, and each form has certain advantages and disadvantages. A firewall's main objective is to enable genuine traffic to pass while blocking harmful requests for it and data packets. Based on their overall structure and mode of operation, several types of firewalls may be categorized. These eight firewall types are listed: Firewalls that filter packets Firewalls with stateful inspection - Circuit-level gateways Proxy firewalls also known as application-level gateways, next-generation firewalls, software firewalls, hardware firewalls, and cloud firewalls.
• FTP: File Transfer Protocol, or FTP, is a network-based client-server protocol for file transfers between clients and servers.
• SMTP: Simple Mail Transfer Protocol, or SMTP, is a standard for sending and receiving electronic mail that specifies the rules and semantics (e-mails).
• DNS: Domain Name System, or DNS, is a naming scheme used for networked devices. It offers assistance in converting domain names to IP addresses.
• Bandwidth: A network's bandwidth is the difference between its upper and lower frequency limits. It is the data transfer rate of a network i.e., measured in bits per second (Bps).
• Node: A node is essentially a location where a connection happens. It might be a piece of hardware or a network.
• TELNET: TELNET offers hosts across the network bi-directional text-oriented services for remote login.

Belady’s anomaly: The phenomenon where increasing the number of page frames causes an increase in the frequency of page faults for a specific memory access pattern is known as Belady's anomaly.

The following page replacement algorithms frequently encounter this phenomenon:

1. First in, first out (FIFO) 
2. Second chance algorithm
3. Random page replacement algorithm

Race Condition: When two or more threads have simultaneous access to common data and attempt to modify it, this is known as a race situation. You cannot predict the sequence in which the threads will attempt to access the shared data since the thread scheduling method might switch between threads at any time. Since both threads are competing to access/alter the data, the outcome of the change in data depends on the thread scheduling method.

Solution to Race Condition: 

• Atomic operations 
• Mutual exclusion using locks 
• Semaphores 

Note: We cannot use a simple flag variable to solve the problem of race condition. 

Paging: Paging is a memory management scheme that eliminates the need for contiguous allocation of physical memory. The process of retrieving processes in the form of pages from the secondary storage into the main memory is known as paging. The basic purpose of paging is to separate each procedure into pages. Additionally, frames will be used to split the main memory. This scheme permits the physical address space of a process to be non – contiguous. Paging avoids external fragmentation. 

TCP: Transmission Control Protocol is a transport layer protocol. Data is reliably and error-free transmitted from the source to the destination computer by use of this connection-oriented protocol. Prior to transmission, a link is formed between the peer entities. TCP segments an incoming byte stream at the transmitting host and gives each segment a unique sequence number. TCP rearranges the segments on the receiving host and notifies the sender that the segments were correctly received. 

UDP: User Datagram Protocol is a transport layer protocol. It is a message-oriented protocol that offers a straightforward unreliable, connectionless service that is not recognised. It is appropriate for applications that do not need the flow control, error control, or sequencing features of TCP. It is used to send little amounts of data where delivery speed is more crucial than delivery precision. 

SCTP: Stream Control Transmission Protocol is a transport layer protocol. It incorporates elements of both TCP and UDP. It provides a dependable, connection-oriented service like TCP while being message-oriented like UDP. It is utilised for Internet telephony. 

Logical Address Space: During the program execution, the CPU creates the logical address. Since the logical address does not physically exist, it is also referred to as a virtual address. The CPU choses to refer to this particular address when attempting to access a physical memory location. The collection of all logical addresses produced by a program's perspective is referred to as the "logical address space.” Memory-Management Unit is a piece of hardware that maps logical addresses to their corresponding physical addresses.

• Swap-Out: Swap-Out refers to the operation of shifting a process from main memory to secondary memory. 
• Swap-In: Swap-In refers to the activity of moving a process from secondary memory to main memory.
• Swap-Space: Swap-space refers to the portion of the disc where processes that have been swapped out are kept. 

Resource Allocation Graph:

A system resource allocation graph is a directed graph that can be used to more precisely define deadlock. The graph consists of a set of edges (set of "E"s) and a set of "V"s or vertices. The vertices V are divided into two groups of nodes, P (P1, P2, P3, etc.), which represent all of the system's processes, and R (R1, R2, R3, etc.), which represent all of the system's resource types.

P -> R represents the direct edge between P and R, meaning that P has asked for an instance of resource type R and is now waiting for that resource.

A resource type R instance has been assigned to the process P, which is indicated by the direct edge from R to P being denoted as R->P.

Here, the resource categories and processes are depicted visually by rectangles and circles, respectively.

A resource allocation graph is not considered to be in a stalemate if it lacks a cycle. Additionally, the system might or might not be in Deadlock if there is a cycle.

Resource request/release life cycle:

The following sequence describes the resource request/release life cycle of OS:

• The process will request the resource. 
• OS demonstrates a clear validation of the process request. 
• The OS will check to see if the resource is available if the request made by the process is legitimate. 
• The resource will be allocated to the process if it is readily accessible. If not, the procedure would have to wait. 
• The process will begin execution if all of the resources needed at the outset are allotted. 
• The process will release the resource after the execution of the process is finished.

Super: 

Whenever a subclass needs to refer to its immediate superclass, it can do so by use of the keyword ‘super’.

Thus, super( ) always refers to the superclass immediately above the calling class.

This is true even in a multileveled hierarchy. 

It is used in variables, methods and constructors. 

Code: 

class Animal { 
public void makeSound() { 
System.out.println("Animal make a sound"); 
} 
} 
 
class Human extends Animal { 
public void makeSound() { 
super.makeSound(); // calls the makeSound() method of the parent class 
System.out.println("Human speaks"); 
} 
} 
 

public class Main { 
public static void main(String[] args) { 
Human A = new Human (); 
A.makeSound(); 
} 
} 

Output: 

Animal make a sound 
Human speaks 

In this example, we have an Animal class and a Human class that extends the Animal class. The Animal class has a makeSound() method that simply prints out a message. The Human class overrides the makeSound() method and first calls the makeSound() method of the parent class using super.makeSound(), and then prints out a message specific to the Human class. 

In C++, the super keyword is not used to refer to the parent class. 

Final: 

When we are declaring a method, adding the modifier final at the beginning can prevent it from being overridden. Methods marked as final are incompatible with overriding. Methods marked as final can occasionally improve performance: Since the compiler "knows" they won't be replaced by a subclass, it is free to inline calls to them.

Sometimes we will want to prevent a class from being inherited. To do this, precede the class declaration with final. Declaring a class as final implicitly declares all of its methods as final, too.

As we might expect, it is not possible to declare a class as both abstract and final since an abstract class is incomplete by itself & relies upon its subclasses to provide complete implementations.

Example: 

final public class MyClassName { 
// Class definition 
} 

This class can no longer be inherited by another class. 

Const: Any moment the const keyword is associated with a method(), variable, pointer variable, or class object, that particular object/method()/variable is not allowed to change the value of its data items. 

Example:

const a = 5;

Now, the value of ‘a’ cannot be changed further.

One of the most frequently posed Computer Science Interview Questions, be ready for it.  

Port: Virtual points where network connections begin and stop are called ports. Ports are controlled by the operating system of a computer and are software-based. Each port is connected to a distinct procedure or service. Using ports, computers may distinguish between different types of communication with ease. For example, although using the same Internet connection, emails travel to a separate port than webpages.

Port Number: Every network-connected device uses a standard set of ports, each of which is given a port number. The majority of ports are set aside for certain protocols, for instance, port 80 is the designated location for all HTTP transmissions. The targeting of certain services or applications inside those devices is made possible via port numbers as opposed to IP addresses, which allow messages to move to and from particular devices. 

There are 65,535 potential port numbers, however not all of them are often used. The following are some of the most popular ports and the corresponding networking protocols: 

• Ports 20 and 21: Protocol for File Transfer across Ports 20 and 21 (FTP). File transfers between a client and a server are done using FTP.  
• Port 22: Secure Shell on port 22 (SSH). One of the various tunnelling technologies used to establish secure network connections is SSH. 
• Port 25: Simple Mail Transfer Protocol (port 25) (SMTP). Email is sent via SMTP. 
• Port 53: Domain Name System, port 53 (DNS). With DNS, users may load websites and apps without having to memorise a huge list of IP addresses. DNS links human-readable domain names to machine-readable IP addresses. 
• Port 80: Hypertext Transfer Protocol, port 80 (HTTP). The protocol known as HTTP is what enables the World Wide Web. 

Deadlock occurs when two or more processes are unable to move forward because they are all reliant on a single event that actually never happens.

Resources assigned to a process are typically not preemptable. This indicates that once a resource is assigned to a process, there is no simple way for the system to remove it from the process unless the process voluntarily gives it up or the system administration kills the process. This might result in a condition known as stalemate.

If there is a stalemate among a group of processes or threads because they are all waiting for resources that are already in use.

Example:
A deadlock in a database occurs when many transactions are waiting for each other to release locks. 

 Deadlock Necessary Condition are as follows:

• Mutual Exclusion: A minimum of one resource must be stored in a non-shareable mode, meaning that only one process may utilise it at a time. This is known as mutual exclusion. The requesting process must wait until the resource has been released if another process needs it sooner.
• Hold and Wait: Minimum of one resource must be held by the process, and it must also be waiting for another resource.
• No Pre-emption: A resource can only be released freely by the process that is holding it once that process has finished its tasks.
• Circular wait: A collection of waiting processes (P0, P1, P2,..., Pn) must exist such that P0 is awaiting for a resource owned by P1, and P1 is awaiting for a resource held by P2….Pn-1 is waiting for resource held by Pn and Pn is waiting for a resource held by P1.

All 4 are the essential for deadlock to occur.

A process can recover from deadlock in the following ways:

• Killing the process:
  • Kill all the processes which are involved in the deadlock.
  • Kill one process after the other based on priority.
• Resource pre-emption: a resource will be pre-empted from the process, which are involved in the deadlock and pre-empted resources will be allocated through some other processes so that they may be possibly of recovering the system from deadlock. If the above condition is followed, it may be possible for the process to undergo starvation.
•  Ostrich Algo: Ignores the deadlock.

Conditional Variable: A synchronisation primitive called a condition variable allows a thread to wait until a certain condition is met.

Conditional variable is used to avoid busy waiting. 

Semaphores: 

A shared integer variable across threads is all that a semaphore is. 

1. Unlike mutex, which only permits multiple threads to access a single shared resource one at a time, allows multiple programme threads to access the finite instance of resources. 
2. We may change the definition of the wait () and signal () semaphore operations to eliminate the necessity for busy waiting. A process must wait when it uses the wait () method and discovers that the semaphore value is not positive. However, the procedure itself causes a car block rather than participating in busy waiting. A process's status is changed to the Waiting state by the block operation, which adds it to a waiting queue connected to the semaphore. The CPU scheduler then assumes control and chooses a different process to run. 
3. When another process performs a signal () action, it should restart all blocked processes that are waiting on semaphore S. A wakeup () action causes the process to resume by moving it from the waiting state to the ready state. After that, the process is added to the ready queue. 

There are 2 types of semaphores: 

Binary semaphore: 

• It is also called mutex locks. 
• The value of it can be 0 or 1. 

Counting semaphore: 

• Can range over an unbounded domain. 
• It is possible to use it to restrict access to a resource that has a limited number of instances. 

Peterson’s solution: Although Peterson's method only works for two processes or threads, it can be utilised to eliminate race conditions. 

Mutex/Locks: By limiting access to critical section to a single thread or process, locks may be utilised to establish mutual exclusion and prevent race conditions. It has starvation of threads with high priority. It can generate deadlock and debugging is a problem. 

Critical Section: The code segment known as the critical section is where processes and threads access and write to shared resources like files and common variables. Any process can be stopped mid-execution since processes and threads run simultaneously.

Concurrency: The simultaneous execution of several instruction sequences is known as concurrency. When many process threads are active at once, it occurs in the operating system. 

Thread Scheduling: A thread's execution is planned according to its priority. Although threads are running within the runtime, the operating system nevertheless allots processor time slices to each thread. 

Time Stamping: A DBMS-created unique identifier called a "time stamp" shows the approximate commencement time of a transaction. Whatever transaction we are carrying out, it records the transaction's beginning time and designates a precise time.

SDLC: The software development life cycle (SDLC) is a method that creates software quickly and at the lowest possible cost. 

Domain Resolution: Domain Resolution is a service that links a domain name to the IP address of a website space so that users of the registered domain name may quickly visit the website. A network location is identified by a site's IP address, which is a numerical address. The domain name is used to identify the web address rather than the IP address to aid memorization. The process of translating domain names into IP addresses is known as domain resolution. The DNS server is responsible for domain name resolution. 

Domain resolution is sometimes referred to as domain pointing, server setup, reverse IP registration, and other similar terms. 

Now, let’s see how domain resolution works: - 

1. A DNS resolver receives a request from a web browser when a user types a domain name into the browser. 
2. DNS resolver translate the domain name into an IP address by querying a DNS server. 
3. The DNS server searches its database for the IP address related to the domain name. If a match is found, it gives the DNS resolver the IP address. 
4. The DNS server will ask other DNS servers until it finds one that has the data if it does not have a record for the domain name. 
5. The DNS resolver gives the web browser the IP address it has received from the DNS server. 
6. The IP address is then used by the web browser to connect to the web server hosting the website associated with the domain name. 

3 Way Handshake:

This may alternatively be thought of as the process through which a TCP connection is made.

TCP uses a technique known as Positive Acknowledgement with Retransmission to guarantee dependable communication (PAR). A segment is the name given to the Protocol Data Unit (PDU) of the transport layer. Now, until it receives an acknowledgement, a device employing PAR must resend the data unit. The receiver discards the segment if the data unit it has received is damaged (it validates the data using the transport layer's checksum feature, which is used for error detection). The data unit for which a positive acknowledgement was not received must thus be resent by the sender. From the foregoing procedure, it is clear that three segments must be sent and received by the sender (client) and receiver (server) in order for a trustworthy TCP connection to be made.

Let’s see how this works:

• Step 1 (SYN): The client transmits a segment with SYN (Synchronize Sequence Number), which notifies the server that the client is likely to start communication and with what sequence number it starts segments with, in the first step of connecting to a server.
• Step 2 (SYN + ACK): The server sends a SYN-ACK signal response to the client request. SYN stands for the sequence number it is likely to start the segments with, and ACK stands for the response of the segment it received.
• Step 3 (ACK): The client acknowledges the server's answer in the last section, and they both establish a secure connection to begin the data transmission.

Different class of Network:

• Class A network: The network is denoted by the first period, and the device inside the network is denoted by the second period. As an illustration, the network is denoted by "203" and the device is denoted by "0.113.112."
• Class B network: Prior to the second phase, the network is indicated by everything. Again, taking the example of 203.0.113.112, "203.0" denotes the network and "113.112" denotes the device connected to that network.
• Class C network: For Class C networks, the network is indicated by everything written before the third period. The Class C network is denoted by "203.0.113" in the same example, while the device is denoted by "112".