Hi guys , just thought about some more topics to explain on this tibco tutorial series , let me know out of below which Tibco topics you would like to me cover first , please suggest if you have other topics in mind. I am now updating this post as most of the topics are already covered but I forgot to update this post.
I have covered difference between routing daemon and Tibco daemon in this part. basically main difference between both of them is the purpose. RVD is mainly for sending and receiving message in same network between two Tibco RV publisher and subscriber while RVRD is used to bridge two different network from two different countries, cities or even continent. Most of the investment banks has RVRD bridging there London to New-york traffic or London to Tokyo traffic. Bandwidth of these WAN networks are very important because if you have just one slow Tibco subscriber that can bring down whole network link between two regions. We had an incident when one of out application is publishing huge amount of updates in MB and on other side we had a GUI on a slow desktop which is keep asking for resend, this problem is often referred as Tibco RV storm and its very serious issue while working with Tibco messaging. One important thing to note here is reliability parameter of tibco rv which actually defines how long publisher should hold the message if set very low can cause frequent retransmission request.
This is one of my favorite Tibco messaging interview question and I have covered this topic keeping both interview and technical knowledge in mind. Since today RV and JMS are two prominent messaging technology where IBM MQ and Tibco EMS both implement Java JMS messaging service while propriety rendezvous messaging is still fastest one available. Anyway main difference between EMS and RV is that EMS is based on JMS while RV is based upon propriety technology. EMS is based upon hub and spoke model which is subject to single point failure while RV is based upon bus model and more reliable as there is no single point of failure. If Daemon dies on any one machine only that machine is affected and rest of machine in the network keeps communicating to each other while in case of EMS or MQ series if Server died than whole infrastructure failed. In general this issue doesn’t happen because most of network has load balanced EMS server so if one goes down other takes it place. On the usage front rendezvous is mostly used for high speed messaging where some degree of data-loss is acceptable while EMS is mostly based upon durable messaging.3. Reliable vs Certified messaging in Tibco RV
If you are going for any Java position in investment bank and if that application uses Tibco messaging than knowledge of certified messaging is must. This is the most asked and yet unanswered question in any Tibco interview. As we know that rendezvous is mostly used for high speed messaging and some form of data-loss is acceptable but there are cases where you just can not afford to loss single message like if you are sending Orders, Trades or Booking messages to middle office system. This kind of messages requires guaranteed delivery which is not possible with reliable mode of rendezvous but no need to disappoint as Tibco provides guaranteed delivery via Certified messaging. It keeps track of delivery of every single messages for every single consumer and holds messages in tibco ledger files until every single consumer acknowledge its receipt. This is the primary messaging solution for highly important messages in trading platforms.
4. Publish Subscribe vs Request Reply messaging.
Tibco rendezvous supports two kinds of messages model one is publish subscribe and other is request reply. In publisher subscriber model there could be one publisher or many subscriber or vice-versa i.e. its one to many or many to one messaging model but request reply is like emails with just one sender and other receiver. Though publisher subscriber is more popular model where server is publisher and clients interested in updates are subscribers, there are messages which is only requested by a single client or only needs to send to single consumer and in those cases request-reply messages are appropriate. I have not covered this topic as separate post but will try to cover it in more detail later.
5. Tibco Hawk
Tibco hawk is an application monitoring tool provided from tibco. You can setup hawk daemon in any Linux machine and then few hawk rules which will than monitor your application health, network connections and any thing which you are interested in. tibco hawk rules mostly specify patterns in log files like errors, exception, full garbage collection and several other parameters. I have covered tibco hawk in detail on this post.
6. Using Tibco API in Java
I have not covered how to use Tibco rendezvous API yet and not sure when can I do that but most of tibco installation comes with sample program in both C++ and Java and you can run those program with any Tibco setup for sending and receiving messages in different tibco rv modes and that is the primary tool I used while learning tibco programming.
7. Using Tibco API in C++
This is mostly an extension of earlier topic but more detailed in terms of Tibco C++ API.
8.Low level implementation of Tibco
This is another interesting topic which I have in my mind but haven’t got chance to cover it.
I have managed to cover few tibco issues and exception as these are day to day problem understanding of where problem lies is important, it could be in your messaging code or it just due to incorrect configuration but most mysterious tibco errors are due to network like every thing looks good but receiver is not receiving messages what can be wrong and than you find out that either daemon is not running, listening to different topic or tibco traffic between those two machines are blocked by network or middle-ware team. I have covered many such scenarios in this post.
10. Tibco Advisory messages
Every Tibco setup publishes advisory messages if something goes wrong or going wrong like Tibco issue advisory messages for retransmission request, for data-loss errors and many other scenarios. I intend to cover this is good detail but let’s see when can I do that. Till than if you come across any scenario which you would like to share than please post as comment.
Finally let me know which topic you would like to cover first in tibco tutorial series, which topic interests you and I will try to cover those, I don’t promise but I will make best efforts to cover them .
Related post:
Fundamental concepts of Tibco RV Messaging
What is Tibco Rendezvous messaging
Top 10 Tibco Rendezvous tips and commands
Tibco Tutorial for beginners
How to use http interface of Tibco daemon for monitoring
Tidak ada komentar:
Posting Komentar