The IBPS SO IT Officer Mains paper rewards depth in Computer Networks more than almost any other subject, because networking questions are concept-heavy and hard to bluff. If you can reason about the layered models, IP addressing, and the common protocols, this section turns into reliable marks. This guide teaches the professional-knowledge core the way it is actually examined.
Computer Networks for IBPS SO IT: start with the layered model
A network is easier to reason about when you split its job into layers, where each layer does one thing and hands its result to the layer below. Data from an application is wrapped, or encapsulated, with a header at each layer on the way down, and unwrapped on the way up at the receiver. Get this idea firmly and most protocol questions stop being memory tests.
The two models you must know are the seven-layer OSI reference model and the practical TCP/IP model. OSI is the teaching standard; TCP/IP is what the internet actually runs on.
The OSI and TCP/IP models: seven layers you must know cold
Learn the OSI layers top to bottom with one function each:
Application: user-facing services (HTTP, SMTP, DNS, FTP).
Presentation: translation, encryption, and compression of data.
Session: establishing, managing, and terminating sessions.
Transport: end-to-end delivery, segmentation, TCP and UDP.
Network: logical addressing and routing, IP lives here.
Data link: framing, MAC addressing, error detection on a link.
Physical: bits on the wire, voltages, connectors, and timing.
The TCP/IP model collapses these into four working layers: Application, Transport, Internet, and Network Access. A common exam trap is asking which OSI layer a device or protocol belongs to. A router works at the network layer, a switch and bridge at the data link layer, and a hub or repeater at the physical layer. Keep that mapping ready.
IP addressing: the arithmetic that carries marks
An IPv4 address is 32 bits, written as four dotted-decimal octets like 192.168.1.10. The address splits into a network part and a host part, and the split is defined by the subnet mask. Classful addressing (A, B, C) is the old fixed split; classless addressing with CIDR notation like /26 is how it is done now.
Work one subnetting problem by hand and the pattern sticks. Suppose you are given 192.168.1.0/24 and must create four equal subnets.
Four subnets need 2 borrowed bits (2 to the power 2 equals 4), so the new prefix is /26.
The mask becomes 255.255.255.192, and the block size is 256 minus 192, which is 64.
The subnets are 192.168.1.0, 192.168.1.64, 192.168.1.128, and 192.168.1.192.
Each subnet has 64 addresses, of which 62 are usable hosts, because the first is the network address and the last is the broadcast address.
That single computation, borrow bits, find the block size, list the ranges, drop the network and broadcast address, answers most addressing questions you will see.
Core protocols: transport and application layer
At the transport layer, TCP is connection-oriented and reliable: it sets up a connection with a three-way handshake (SYN, SYN-ACK, ACK), numbers bytes, acknowledges them, and retransmits losses. UDP is connectionless and lightweight, with no handshake and no delivery guarantee, which is why it suits DNS lookups, streaming, and voice. Knowing which application uses which, and why, is a frequent question.
At the application layer, keep a short table in your head: HTTP and HTTPS for the web, DNS to resolve names to addresses, SMTP, POP3 and IMAP for email, FTP for file transfer, and DHCP to hand out addresses automatically. ARP maps an IP address to a MAC address on the local link, and it sits right at the boundary of the network and data link layers.
A couple of related ideas round out the topic. IPv6 uses 128-bit addresses written in hexadecimal, introduced because the 32-bit IPv4 space is exhausted. NAT (Network Address Translation) lets many private hosts share one public address, and a port number identifies the specific application a segment is meant for, which is how one machine runs many network services at once.
Network security basics for the IT Officer role
Because the post is a banking IT role, security shows up in the professional-knowledge paper. Cover the essentials confidently:
Symmetric encryption uses one shared key (AES, DES); asymmetric encryption uses a public and private key pair (RSA).
Hashing (SHA family) gives a fixed-length fingerprint used for integrity, not for reversible secrecy.
A firewall filters traffic by rules; TLS secures HTTP into HTTPS using a mix of asymmetric key exchange and symmetric bulk encryption.
Recognise the common attack names: denial of service, man-in-the-middle, phishing, and spoofing.
How Computer Networks is tested in the IBPS SO IT Mains
Professional Knowledge in the Mains draws on this exact syllabus, and networking is one of its densest areas. Our published question bank carries over 2,100 Computer Networks questions, so there is no shortage of practice on layered models, subnetting, and protocol identification.
The exam favours precise conceptual questions: which layer does a protocol belong to, what does a given TCP flag do, how many usable hosts in a /27, which protocol is connectionless. Treat every official specific, the number of questions, the sectional timing, the marks, and any negative marking, as something to confirm on the current notification at ibps.in, because those numbers can shift between cycles. The concepts here do not shift, and that is what you should bank on.
Your next step
Computer Networks is learned by working problems, not by reading summaries. Build the subject properly and then drill it.
The core paper for this stage is prepared in the IBPS SO IT Officer Mains course, which frames networking for the banking IT role.
If you want all three stages sequenced together, the IBPS SO IT Officer Scale-1 bundle covers Prelims, Mains, and Interview in one place.
For layer-by-layer depth, work the OSI model learn module and practise addressing on the subnetting learn module.
More banking-IT preparation sits under the Banking and Insurance category.
Learn the seven layers cold, hand-solve two subnetting problems, and keep TCP and UDP cleanly apart. Do that and Computer Networks becomes a section you count on rather than fear.
