What needs improvement with IBM WebSphere Application Server?

Sometimes, I feel WebSphere runs a bit slow. It might be loading unnecessary libraries, impacting its performance compared to other application servers. So, better performance is my main request.

The solution consumes hardware.

The main issue we faced was its limited compatibility with non-Java technologies, which can result in difficulty detecting potential bugs and requiring additional integration efforts. Third-party support can be challenging as well, as the IBM WebSphere Application Server is known for its limited community and lack of an open-resource library.

The cost of the solution is high and has room for improvement. The installation has room for improvement.

IBM WebSphere Application Server hasn't changed much. It's still a heavyweight for any company compared to what you get. Unless your code base is deeply linked with it, I don't think it's a great idea to go with this solution. The current trend is toward modularity and containerization, and given the product's requirements, containerization will be difficult. There is a memory requirement as well. With the enterprise version, there are features we don't need, but if we did, they are available with the Apache package. You can easily find people who know Tomcat, but you find very few people who deeply know WebSphere. There is a steep learning curve as well. There is hindrance to testing, that is, IBM WebSphere Application Server will take 30 seconds to start, whereas Tomcat will take 2 seconds without any applications. That's a big problem. The installation is very complex. The licensing cost also is 1,000 of euros for a 30-year table. The product has so many problems and needs so many patches. It does not take the standard Java Development Kit; you need IBM 1 even though Java is supposed to be portable. I think there is barely any use case for IBM WebSphere Application Server.

WebSphere Application Server doesn't have an automated deployment option, forcing us to use third-party tools like Jenkins UCD and Palo Automated Deployment. In the next release, IBM should invest in automated deployment instead of complex building integration with different tools.

Most of my clients are quite happy with the WebSphere application, but I know that some are changing direction and the current trend is to move to Liberty because of the portability of its cloud and its Kubernetes, which containerize the application. Since most of the application vendors our customers use also offer the application on Liberty, I think they're probably going to stop enhancing the WebSphere Application Server and instead concentrate on Liberty.

One of the things that we have struggled with is understanding what's happening inside the covers when we're running a JVM. When we run into memory or locking issues, we resort to using third-party tools. However, it would be preferable to have native tools for debugging this type of problem. It's not bad lately in terms of performance. WebSphere has always had some performance issues, at least until about three or four years ago. But it's getting better. I guess, aside from the product's complexity, I think it's pretty good for what it's billed as an enterprise application server.

The licensing could be improved, and I would like it to give the longevity of the lifespan of the visions. In the next release, I would like to be able to download and extract the files so that I can just use my application server.

In spite of the solution's robustness, it is expensive and a bit difficult to support, which is why companies nowadays tend to use more lightweight products such as Tomcat or cloud versions of the products. We are also moving to cloud versions and have a huge installation of IBM WebSphere as a legacy system. Probably, in two or three years we will migrate to cloud versions. The initial setup is a bit complex, although easy management is possible once one has set up the environment.

The solution could improve the integration.

It is very tough to get developers. It is not open, so there is a shortage of its knowledge in the industry. We have to get freshers and train them. We can't just go out there in the market and get developers who are proficient in IIB. I have attended several boot camps on AI and other products of IBM. Similar to what IBM is doing with big data and AI, IBM should open up IIB so that there is a lot more knowledge. They should open up the WebSphere application so that there is a lot more knowledge. The business logic side of it is sort of missing in the sense that if I want to track and measure velocity, it is not really available. You have to buy another application and embark on a separate implementation. Instead of having different licensing, IBM DataPower should be integrated with WebSphere. It will allow us to build the business layer and rules a lot more efficiently, rather than developing rules within the application. It would be good if we can set up the business layer through parametrization rather than development. IBM DataPower has the business rule and the controls, and if it can be integrated, it would be fantastic. It will help the application in working better in terms of security features and business logic. If you're going to use it for open banking, you will be able to monitor velocity on the total pricing. Its price is a bit expensive. They should improve its price to compete with other applications that are out there, which we are also exploring.

I'm not certain if the WebSphere solution was deployed by IBM. There are a lot of complexities in how the solution was actually built and deployed, which means troubleshooting on management for us is pretty difficult. One of the biggest issues that we've had is there are certain features that we required that were hardcoded into the solution itself. When you manage them for making any architectural or solution changes, it becomes very difficult and near impossible to do. With respect to that, we tried to change the SSL certificate that would be in use, and because of how we tried to change the SSL certificate, we tried to change the DNS mover that it was pointing to. There were hardcoded elements in the solution that didn't make it very easy for us. At the end of the day, we just kind-of renewed whatever services that we had already ongoing with it, which was a duplicate payment with what we had from other sources. We couldn't take advantage of the shared resources that we had before. We now have to maintain it as an isolated instance. Based on the field and based on the build that was provided, we've noticed a lot of constraints in terms of the performance now. Due to GDPR and other issues, not everybody is able to utilize cloud services. That's something that people need to be aware of. The company needs to be clear on the business use case and how they need to maintain compliance with its policies and regulations. Some of the feature sets that we found a little lacking in this particular solution. By now they've probably changed the ability to embed and utilize the rich media content and web presence. Our site is basically little image JPEGs, and that's it. We have low embedded video. We have low dynamic speech response for mobile viewing, we have low integration or extension for mobile apps. We have low integration as well as for dynamic content of bits from other sites. For some of our clients who wish to display information on our website, we actually have to lift the content, reform our tips, and recreate it into the content management engine.

The availability of the solution needs improvement. The product should be more enhanced and responsive. We need to have the capabilities to customize it a bit more. Their pricing is always too high. The user interface isn't too impressive. The stability of the solution could be much better. The solution's technical support could be improved. They can be too technical and too detailed.

Cloud service offering should be improved because the future is in the utilization of the PaaS provided by principal cloud providers; I mean that it would be very useful to have the possibility to use WebSphere as a service like many other tools has already done.

Some things are very difficult to do, so the interface and usage could be more intuitive for those. The main complaint that my customers have about this product is the price.

It should be able to serve more concurrent requests like Oracle. Oracle has more powerful stability, availability, and real-time serving. WebSphere is not light enough to implement high available applications like gateways. But WebSphere has more configuration abilities and customizability.

Nowadays the industry is moving towards a more open-source operating framework. The cost factor is huge. It's very pricey compared to other open-source stacks. In the future, we'll deal with the IBM Stack so we might move to a compact server and other open-source alternatives which are comparatively less in terms of cost. They should make the solution more lightweight and not bundle everything into a single product. The solution would benefit from having a different licensing model.

In the next release of this solution, I would like to see support for the Arabic language. Better and easier integration with other solutions, such as DB2, would be an improvement.

The computing resources required to run WAS could be reduced so that we can use a smaller virtual machine to run the application. I would like WAS administration tasks to be completed via scripts, rather than GUI.