Tibco BW5 is a great productivity tool, but the new version, BW6, is not.
Updated April 16, 2018

Tibco BW5 is a great productivity tool, but the new version, BW6, is not.

Benjamin Warrick | TrustRadius Reviewer
Score 5 out of 10
Vetted Review
Verified User

Overall Satisfaction with TIBCO BusinessWorks

We use Tibco Business Works to integrate and automate global and local enterprise systems -- ERPs, databases, SaaS, REST and SOAP APIs, and legacy systems of numerous types. Our organization has over 200 global locations and our work includes many critical business processes. We use a service oriented architecture (SOA) and have become a little famous in our organization for speed and quality.
  • TIBCO Business Works does field mapping well. I often have projects that include ERPs like SAP, which can often use large and complex XML schema. BW makes it easy to map fields between these cumbersome documents, loop over repeating nodes, apply xpath... I think this is where BW has the biggest advantage over other tools.
  • BW also makes it easy to establish logical flow. When it's just code on a page, you have to rely on UML diagrams. BW offers an icon based designer to design the logical flow, which is pretty nice and intuitive.
  • TIBCO BW runs very efficiently. People are often surprised at how quickly our systems move, especially when components like the database adapter are used. We have many integrations quickly moving large amount of data.
  • Training developers new to Tibco BW5 easy. This might be my favorite thing about it these days. If someone already has experience with app development, or is otherwise familiar with business processes, sql, xml, ems, etc., BW5 is very easy to be productive on.
  • General development time on BW5 is quick. From a business perspective, I would say this is the strongest advantage. Creating new processes and making adjustments is fast and reliable -- not prone to creating new issues.
  • When discussing needed improvements to BW, we have to make a clear discrepancy between BW5 and BW6. BW5 is stable, tried-and-true, and easy to train people on. BW6 is the total opposite. Also, BW6 offers no significant advantage over BW5. I began using BW6 almost two years ago now. Our original intention was to make use of its REST activities first thing, then transition all our processes to BW6. Today I still only use it for REST, but I don't use the REST activities because they can't do complex queries. I have to use HTTP activities... We also have NO plans to move anything other than REST interfaces to BW6.
  • The main issue with BW6 is the dreadful bugs. It's a sophisticated system but very immature, so it's loaded with dangerous surprises. I even took the BW6 bootcamp class, intending to get a good start. But the class was mostly worthless because the instructor couldn't even run the example solutions. We just fumbled though numerous introductions to features. The best advice I can give is, first learn BW5. Then, if you really want to learn BW6, go ahead, but don't use it for any projects with critical deadlines. Also, push to the repo often, and understand that your workspaces are basically disposable. Things go wrong that just aren't worth all the XML editing to fix, so it's better to just backup an hour or two and start again. Also, delete and recreate if anything starts to act funny. When the designer crashes, start it up, shut it down, then start it up again. Also, you're not alone. Everyone knows this thing is buggy.
  • The BW6 designer is built on Eclipse. I was optimistic about this, and anticipated a more efficient interface compared to BW5's designer, which is clunky. BW6 designer looks more slick, but it offers no real advantages over BW5, and is heavy and unstable. It crashes often, is painfully slow, corrupts project files, and otherwise requires a lot more time than BW5 for even simple tasks.
  • Testing is more difficult on BW6 than on BW5. Setting up a whole integration and running it locally is an easy task with BW5. Things run quickly. It's easy to see the input and out put data from each process. Errors are quickly evident. BW6 -- just getting it to build and run can be difficult. The big issue is that there are numerous run time hazards that BW6 will not indicate after compilation. You just have to run the thing and sift through terminal output. Here's a tip. It always say every process is imparted. The gobs of red text that poor out, that's normal and OK. Neither of these sections are describing your error. Your error will be in black text somewhere.
  • Once you get your integration running, examining the input and output is challenging. The designer will take your focus for various reasons. For example, if you have http debugging enabled, you simply can't look at anything else. The focus is repeatedly returned to the standard output terminal. Or, even with minimal debugging and nothing trying to grab focus, just digging up the data can be difficult, especially if you're dealing with large objects. You can't click on an activity to see its output, as with BW5. You have to sift through the nested lists of processes and down into activities. It's actually hard to even explain how to see input/output without an image.... But you can't really even see the data from within the designer itself. You have to copy it out and paste it into a text editor.
  • Online training
  • In-person training
  • Self-taught
For the old version of Business Works, version 5, we had a consultant help us with getting started. He was very helpful and BW5 was relatively easy to learn and deploy. Most things we could learn on our own. For BW6 we tried Tibco's online training. It was not effective because the lessons built upon each other, but there were so many software bugs it was not possible to get each lesson working. The lessons also consisted of mild introductions to numerous concepts, which could have been accomplished just as quickly by reading the documentation, and offered no advantage over simply skimming the docs.
  • BusinessWorks 5 allowed us to make a good name for our new team. We were able to create solutions quickly, and they ran reliably. We were not only fast and reliable, we were also able to accomplish projects that had failed in the past, mostly due to a lack of software developers. With TIBCO, we were able to make it much easier for the various teams to share data. BW6 has compromised our good name so far. It often takes a very long time to develop even a simple interface, and sudden strand issues have halted our production instances.
  • BW5 is expensive but worth the cost. Two people, already well versed in the business processes and infrastructure of an organization, can easily and quickly create interfaces between systems. The technical work is probably 10% of the effort, the rest business process discovery and testing. BW6, again, is a real pain. We only use it for REST calls. BW6 is not worth the cost.
  • Training a new developer on BW5 is relatively easy. For example, it only took me 12 hours to train a Java developer to use BW5 productively, able to work on actual projects. My boss and I also self learned BW5 over the course of running projects, and found it straight forward. We did no formal training. BW6 is again opposite. The issue is the Eclipse based designer. Eclipse is a good tool, but the BW6 designer is very immature and suffers from many frustrating bugs -- some that even corrupt the work-space or crash the designer. The biggest challenge with BW6 is getting around these bugs. For this reason, even after having it for well over a year, I have not started to train any new developers on BW6. It would simply be a waste of time. They would not be productive for months and would constantly need my help with complicated issues and vague errors. It's just better to do the REST interfaces myself. BW6 is not worth the time cost of training.
The big advantage of BW5 is the icon based designer. Essentially, TIBCO has done all the very difficult and very easy stuff for you. What's left is designing the business process logic and interface settings. The mapping tool is also very powerful. For example, it's relatively easy to map fields from an SAP idoc to a totally different format, like a set of database tables of different ERP. You can even use Xpath for more complex logic. As far as efficiency and speed goes, BW5 is the best system I've seen, and I can't actually think of any way to improve it significantly.

BW6 offers essentially the same business process design and data mapping/translation abilities. The issue is that BW6 is built on Eclipse and had a lot of bugs still. So development on BW6 designer can be unpredictable with regard to time and efficiency. I would rather it not be part of our tool set, actually. It's proven more of a liability that has made our team look bad several times now.
Data mapping is very good with BW5 and BW6. But like I've been saying, BW6 isn't worth the money or significant learning time. BW5 is great. I've implemented some very complex data mappings, where some fields are based on other fields, etc., and I can always solve it with BW5. A lot of integrations we've made would probably not be possible without BW5, because of the complexity. It's also very easy to look at the complex map to determine the logic and solve issues or ask questions. It's often the case that business requests numerous changes to this logic, and a month later has questions about the logic they requested. All I need to do is open the project and examine the map, or even step a message through to see exactly.
The advantage of Tibco BW5 is that it's easy to watch the data flow through the process. If there is an issue with the database query, it runs fine until it hits that process, and fails with the error message. It's very easy to run tests and analyse the data going through. You can even run multiple instances of the designer and watch data pass between them.

With BW6 it's a more cumbersome process and you'd better have plenty of memory on your machine and a lot of patience waiting for everything to start up. Good eyes is also essential because BW6 dumps a huge amount of standard output, which often includes run-time errors that may or may not be actual issues.
Mulesoft is good for creating APIs.
TIBCO BusinessWorks 5 is good for any situation where you need to move a lot of data between different systems. It's fast once deployed and development is fast too. BusinessWorks 6 is still very immature. I wish we would have waited at least another year before getting tangled up with it. I think I've made five necessary upgrades in that time, and I'm currently behind by at least one.