Compare & Contrast: Hyland's Process Automation Platforms
Hyland’s annual customer conference for 2024, CommunityLIVE, was held on September 23-26. They unveiled their vision for the future of their company which was centered around federated content services and artificial intelligence, but also included were demos featuring Hyland Experience Automate (HxA). HxA is Hyland’s next-generation Business Process Model Notation (BPMN) process automation platform. People who are familiar with Alfresco Process Automation (APA) or even Alfresco Process Services (APS) may notice a lot of similarities within HxA.
Why is that? It’s because all three products are built on the Activiti BPMN engine, but each have their own difference that make them unique. So, if you are already an APS or an APA customer, you may be wondering what the differences are and should you consider switching to HxA? Here at ClearCadence, we have experience in all three platforms and are here to help navigate you through the differences between each, focusing on their core strengths.
Alfresco Process Services (APS)
Let’s start with Alfresco Process Services (APS). This is the only on-premise, Activiti based platform offered by Hyland. The Activiti version that is the foundation for APS is 6.9 whereas the latest version of Activiti is 8.7. So while APS is a few versions behind, it is still a powerful BPMN platform providing extensibility via REST APIs and connectors. APS is also a stand-alone platform that does not have to be tied to a content repository, but it comes with connectors to Alfresco Content Services (ACS). There you can perform advanced functions like updating metadata and moving documents. APS also provides connectivity to Google Drive and Box, however integrations with these systems are only to publish content.
Being an on-prem platform, APS provides a customer more control over the system internals and databases than they would with APA or HxA, for they are hosted, cloud-based environments. You can take advantage of configuring performance in APS by tweaking database indexes and/or splitting off certain tables to a more distributed environment.
The process modeler in APS uses the same Activiti based notation used in the other two offerings, but structured differently. It provides the Activiti functions like Start Events, Activities (user, script, email, Java code, and REST tasks along with tasks that support Camel and Mule integrations), gateways, and other functions.
APS also has boundary events, but some of their functionality falls short of what is available in APA and HxA. For example, the boundary error event will work for script tasks when an error occurs, but not for REST call errors. Since tasks can get stuck in a REST call because of an error, the process instance cannot be recovered. Overcoming this is part of a sound development and testing strategy to make sure you have controls to either prevent the situation from occurring or to have a back-out plan. Unforeseen production issues DO happen, and you could be in a situation where a process instance must be restarted because it got stuck, with no ability to recover or catch the error. There are ways around this even without boundary error events and something we will explore in a future blog post.
Alfresco Process Automation (APA)
Now let’s look at Alfresco Process Automation (APA). This is Hyland’s first cloud/PaaS offering that is based on Activiti 7x. APA comes bundled with the cloud version of Alfresco Content Services (ACS). Because it is tied to ACS, you cannot get APA as a stand-alone BPMN platform. Whether you choose to take advantage and utilize the ACS features or not, APA is a part of the ACS cloud environment (although APA may need to be activated via Hyland support).
The tighter APA integration with ACS brings some significant benefits including a more direct approach to ACS functions like Move or Copy. In APS, you must utilize the Call Alfresco Action task for these, but in APA they are defined as independent actions. Additionally, APA provides Triggers which include the ability to monitor certain folders in ACS, and when new content is added, a process instance in APA is created. For on-prem implementations this is usually handled on the ACS side using Folder Rules and custom code but with APA, it’s built into the Trigger functionality.
The left side menu for APA provides all the tools for building the entire application, including multiple processes and forms.
In APS, these are separate libraries accessed across the top menus, and while getting into each to edit them are essentially the same, the APA layout provides an easier view into what processes, forms and other configurations are a part of the overall application. For processes, you can even add groupings to make it clear the function of the process (like Main Process, Sub Process, or Utility).
One big deviation from APS is that APA cannot support Java-based service tasks, whereas with APS you can create a Java delegate to handle more complex code, and release it when APS is deployed. In APA, however, it is not possible to upload your custom Java code. This means code would have to be written in Script tasks or utilize REST services to execute the code and pass back resulting values (or utilizing AWS Lambda to run your custom code. More on that below). Setting up these REST calls in APA is just a matter of configuration, so the effort is straightforward.
To make APA more like a Java IDE, JavaScripts are stored as independent objects that can be referenced within Script tasks added to the process map. These are accessible from the left-side menu navigation (one small complaint here is that the scripts used to be displayed alphabetically which made it easier to get to a script in question, but recent updates have removed that feature).
Another nice feature with the cloud-based Script editor is the Simulator function that allows scripts to be tested without having to publish the entire process. Values for input variables are added in the Script Properties panel, and by clicking the simulate button, the output variable values will be displayed (assuming the script has proper syntax and logic). It should also be noted that unlike APS, APA does not support Groovy Script, just JavaScript.
Looking at the integrations provided by APA, it is obvious Hyland is taking full advantage of being in the cloud and hosted within Amazon Web Services (AWS). There are direct integrations to AWS features such as Lambda (to run your custom code), Comprehend (machine learning for content), Rekognition (image and video analysis), Transcribe (speech/language recognition/translation) and Textract (extracting text and data from content). Additionally, there are connectors for Salesforce, Slack, Twilio, Docusign, Microsoft Teams, and Calendar. All the integration points have their limits, but according to Hyland new features and functionality with existing connectors and the addition of new connectors will be forthcoming.
Hyland Experience Automate (HxA)
This leads us to the latest BPMN offering from Hyland, Hyland Experience Automate (HxA). HxA is also a cloud-native offering that includes all the APA features we discussed above, plus the look and feel is much like what we see in APA. This is a natural evolution since HxA is also based on Activiti. The left side menu is pretty much identical to APA but with a few extra features we will get to in a minute.
Unlike APA, HxA is a stand-alone BPMN process engine, meaning it does not have to be tied to any content repository, if a customer does not have a need to use that type of feature. However, if they do, HxA provides not only its own, out-of-the-box internal content repository (called Hyland Experience Repository – HxPR) it will also integrate with any of the Hyland ECM offerings: Perceptive, Alfresco, Nuxeo, and OnBase. Hyland refers to this group as "PANO".
This means customers with these content repositories (and let’s face it, many large companies end up having more than one in their organization) can utilize HxA to put structured workflows around their content to enhance indexing, application processing, accounts payable, human resource onboarding/offboarding, and many other types of applications. With functions like reading and creating and updating document metadata, the basics are covered across all of the PANO repositories. Some features are different than others, but Hyland plans to increase the functionality for each of these repositories and add additional connectors to other repositories like SharePoint, Documentum, and FileNet. This is all part of their grander vision of the future for federated services and content, i.e. the Content Innovation Cloud:
But, again, a customer doesn’t have to have ANY of these repositories, and can instead utilize the out-of-the-box HxPR functions. These functions include all the basics for content management: create content, update metadata, move content, etc. You can also set up content types, schemas, renditions and security policies from the left side navigation menu. All, right out of the box.
Obviously, the biggest benefit to both APA and HxA is that they are managed cloud services which means a company doesn’t have to buy servers or pay for people to maintain them. Additionally, companies don’t have to worry about major upgrades, because as each incremental version release of the software comes out, the environment will be automatically updated. This means you are always up to date with any new features or bug fixes without having to apply them yourself.
The downside however is little to no control over the underlying environment, which includes no access to backend databases and limited access to log files. For example, you can get to the last 4000 or so lines of a log file, but that may not go far enough back to cover your needs. Hyland is addressing this for APA with a tool that will provide more access to logs and information about the systems. This will be a cloud based feature that will be released in the coming months. Presumably, HxA will provide the same, but no specific announcement has been made.
To summarize the major differences between the Hyland BPMN offerings, here is a table that lays it out:
Depending on your needs, chances are Hyland has the BPMN platform that's right for you. Need to have control over your databases and configurations? APS is the right fit. Already have ACS in the cloud? APA is probably there and ready to be used. Working with Nuxeo AND OnBase and need a single workflow for both? HxA is your answer.
Over the coming months, ClearCadence will be providing more blogs highlighting specific features, tips, and perspectives relating to all three of these products as well as the newer innovations Hyland is putting out like Hyland Experience Insight and Hyland Intelligent Document Processing.
There is a lot to like about what currently exists at Hyland and what will be coming out, but with so much new information, it could lead to confusion. ClearCadence is here help you assess and decide the right Hyland platform, architecture, components, and design to best meet your process automation challenges, either in the cloud or on-premise.
Kevin Beddingfield is a long-time BPM architect with a specific focus on automating business processes - from training on BPM concepts to designing, developing and implementing end-to-end BPM solutions for Fortune 1000 clients.
ClearCadence has a long track record of assisting customers with analyzing, planning, designing, and implementing solutions. Visit the link below for more information about our organization.
Comments