---
title: "Short note: TIB/Rendezvous"
subject: "Distributed System"
chapter: 4
key: "semester_7/distributed_system:rendezvous-note"
source: https://www.bctengineer.com/q/semester_7/distributed_system/rendezvous-note
site: BCT Engineering
---

# Short note: TIB/Rendezvous

Distributed System · Chapter 4: Distributed Heterogeneous Applications and CORBA

- Asked in: 2075 Asoj, 3 marks; 2071 Chaitra, 3 marks (asked 2 times)
- Answer checked by Claude Opus 5.5 (high effort) on 2026-10-14
- Page: https://www.bctengineer.com/q/semester_7/distributed_system/rendezvous-note
- Whole chapter: https://www.bctengineer.com/md/semester_7/distributed_system/chapter-4

## Question

Write a short note on TIB/Rendezvous (Rendezvous concept and implementation).

## Short answer

TIB/Rendezvous is a publish-subscribe middleware using subject-based addressing. Publishers and subscribers communicate anonymously via a daemon (rvd) using multicast. Reliability is achieved with sequence numbers and retransmission; certified delivery guarantees message receipt. It provides low-latency, many-to-many communication suitable for real-time systems like financial trading.

## Answer

**TIB/Rendezvous** is a message-oriented middleware product from TIBCO for building distributed applications using publish-subscribe, request-reply and multicast messaging. It is used in financial trading and other real-time systems where many programs must exchange data rapidly and reliably.

#### Rendezvous concept
- Based on **subject-based addressing**: a message carries a hierarchical subject name such as `STOCK.NEPSE.NABIL.PRICE`. A publisher sends to a subject; any program that subscribes to it (wildcards allowed) receives the message. Sender and receiver do not know each other's address, so they are loosely coupled.
- Communication is **anonymous and many-to-many**; supported styles: publish-subscribe (one to many), request-reply, and certified delivery for guaranteed messages.
- Messages are self-describing, with typed fields, so different languages and platforms can exchange them.
- Reliable multicast over the LAN gives high throughput and low latency.

#### Implementation
```text
 App --> RV API --> rvd daemon ==== network ====> rvd --> App
 (publish)  (client library)   (multicast/UDP)   (subscribe)
```
1. Each machine runs the **Rendezvous daemon (rvd)**. Applications use the RV client library to talk to the local daemon.
2. The daemon sends a published message by UDP broadcast/multicast on the network, and filters incoming messages by subject for local subscribers.
3. **Reliability**: messages carry sequence numbers. A receiving daemon that finds a gap sends a negative acknowledgement, and the sending daemon, which keeps recent messages for a short reliability interval (60 s by default), retransmits them. Order is kept without a central server.
4. **Rendezvous router daemon (rvrd)** passes messages across subnets and WANs; **Certified delivery** records messages in a ledger (usually a disk file) and retransmits them until every subscriber acknowledges them.
5. **Fault tolerance** uses a set of replicated programs, with one active member and standby members taking over on failure.

## Check yourself

The answer's key points as multiple-choice questions, with the correct option and why.

1. How does TIB/Rendezvous address messages?
   A. By network address
   B. By hierarchical subject names that subscribers listen to
   C. By filename
   D. By process ID
   Answer: B. Publishers and subscribers don't know each other, so they are loosely coupled. (Key point: Subject-based addressing)

2. What does the rvd daemon do?
   A. Stores messages permanently
   B. Compiles IDL
   C. Sends published messages and filters incoming ones by subject for local subscribers
   D. Routes across WANs only
   Answer: C. Each machine runs rvd; apps use the client library to talk to it. (Key point: Rendezvous daemon)

3. How is reliability achieved without a central server?
   A. Sequence numbers, negative acknowledgements and retransmission from a ledger
   B. Two-phase commit
   C. A central broker
   D. TCP only
   Answer: A. rvrd crosses subnets; certified delivery stores messages until acknowledged. (Key point: Reliability)

## Related in other subjects

- [DSM concept; planning and implementation steps](https://www.bctengineer.com/question.md?key=semester_7%2Futilization_of_electrical_energy%3Adsm-concept-planning-implementation) — Utilization of Electrical Energy, chapter 6
- [Interface vs implementation](https://www.bctengineer.com/question.md?key=semester_6%2Fobject_oriented_analysis_and_design%3Ainterface-vs-implementation) — Object Oriented Analysis and Design, chapter 4, asked 2×
- [Intranet implementation procedure and VPN](https://www.bctengineer.com/question.md?key=semester_8%2Finternet_and_intranet%3Aintranet-implementation-procedure-vpn) — Internet and Intranet, chapter 6