IBM Workload Automation Reviews

• Seamless integration with different applications like Oracle, SAP, Hadoop, JMS etc.
• Web console with single sign on feature.
• TWS as a tool has features that integrate with different applications and backend technologies, as mentioned, to run the jobs more native to the platform. Say, for example, if I would like to trigger a SAP job from TWS scheduler, I can specify most of the parameters from the TWS forms. As an enterprise scheduler, I have more control over most of the enterprise applications to which it is connected to.

Improved migration/upgrade features helped us to reduce the time of upgrade.

TWS is evolving from V 8.3 to V9.3 on the features and no major changes on the architecture. The user experience side of the console is being improved in all these versions, however, the console is not very fast as expected.

It would be good if the TWS consoles and front end reporting explores different users like business, technology and application teams and seamless and faster experience like mobile apps.

If these supported applications would be provided without additional licenses, it would be good.

Yes - UNIX installation, we encountered issues due to not having sudo root access.

In some windows installations, we faced issues with a gap in the path name, ex: "C:\Program Files\IBM".

No.

No.

3 out of 5 - in cases of high severity production issues, the SLA for PMRs (Problem Management Records) is 2 hours. However, from a business standpoint, I would not be happy to have my business down for 2-3 hours.

No

Initial setup for TWS 9.2, I find it complex, since there are number of components needed to be installed with no clear understanding of why those components are used and what is a pre-req for what.

However, I did find the installation for TWS 8.6 to be pretty straightforward.

In-house.

Seems very low and rate of realization is too slow for simple networks. However, good for high investment and complex scenarios.

There is a general perception that pricing and license costs are too high with the conventional model, like per CPU based prices.

However, IBM is trying to use the per job pricing in the cloud model.

No

Yes, IBM has to review the non-cloud low pricing models and invest in architecture based revolutions.


Additional integration to IBM Watson for analytics would be more helpful.

Workload automation (WLA) these days is no static business. It’s all about running the right workload sequence, at the right time, often triggered by a variety of possible (combination of) events. For instance, we use this principle for running a daily Oracle backup workload batch, which (per backup) involves different systems that in real time have to exchange certain real-time information to be able to successfully end and register the backup in RMAN. Therefore, we use complex event rules that monitor events during the backup process, take care of passing the desired info from one system to the other, and dynamically submit certain jobs that cannot not be defined in advance.

ERP (SAP) connectivity: Thanks to this technology, we were able to fully integrate SAP OS, ABAP and BI (proceschain) workloads with non-SAP workloads, so that a complete business process involving SAP and non-SAP systems could be modelled within one TWS batch. Example: a non-SAP MFT job delivering data @ SAP-PI; in-time PI channel trigger modifies the data for next job; SAP ABAP job works the data into the system involved.

Conditional Branching: WLA often is about critical paths that determine if a batch will be able to end just in time. For instance, C.B. makes it possible to dynamically decide if parts of a predefined batch can or should run in sequence or parallel, depending on the outcome of a certain measured condition, therefore able to meet end-time requirements, even when parts of the batch encounter delay.

HA-DR implementation: No system engineers are required to realize value from these resources @ the application level. TWS makes it possible to use resources, provided at the Application Management level, to meet business requirements for high availability and disaster recovery (HA-DR). We use these resources a lot, not only in case of disaster (e.g., scheduling plan breakdown on master manager; switch to backup-master manager management), but also when TWS management systems in one DC need technical maintenance. In that scenario, we simply switch all our TWS management activities to backup counterparts in another DC, that silently have received up-to-date data from message broadcasts within the TWS network.

The tool made it possible to automate the technical workflow within complex business process models, within a heterogeneous network hierarchy, consisting of fault-tolerant distributed agents, managed by (master-)management agents.

The tool made it possible to fully automate actual job-workflows that represent complex end-to-end business process models (BPM’s). Those workflows run on systems that reside in a distributed, heterogeneous, logical agent-network (fault-tolerant, ‘extended’, SAP R/3, SAP BI and even cloud/dynamic (e.g. Salesforce, Microsoft SQL) brooker agents).

This agent-network itself is managed by (master-)management fault-tolerant agents that periodically provide all the agents in the network with a static (predefined) scheduling plan for a certain future period of time, and also house a so called event-processor, that is able to have the master-management agent submit just-in-time defined workload into the actual running BPM, as a result from detected events within the network.

Today, TWS (TWA) has evolved into version 9.x and the product is now also available as a cloud-provided service (the management parts, from IBM SmartCloud).

I think that is a good and modern development, but the first v9.x releases (and maybe also the latest, I don’t know) lacked the event-driver WLA functionality that was already available for years in older on-prem versions of the product.

That doesn’t help IBM have customers migrate from older on-prem versions to these modern cloud-provided versions, when they would like to migrate.

IBM should have its cloud service deliver at least the same functionalities as their on-premise service has been delivering for years.

I have used it for more than 10 years.

No, but it it has been a joint effort with IBM Tivoli L3 support, PostNL IT RHEL and Windows administrators and a dedicated WLA coördinator / TWS administartor.

I have not encountered any stability issues at all. No other IT infrastructure @ our company has proven to be as stable as our TWS v8.5.1/8.6 network and all its components.

I have not encountered any scalability issues at all.

Very high. Excellent support, dedicated to their customer

I receive very good support from a real product specialist.

Over the past 10 years, I frequently contacted IBM to have a ‘PMR’ registered concerning bugs, problems during admin operations, advice on best-practice batch modelling and in one case, for detailed help during a complex scheduling-plan recovery procedure that did not lead to the desired result. I encountered my own knowledge limitations, but was not given the time to acquire that extra knowledge first.

We did not previously use a different solution. We started implementing a WLA infrastructure in 1996 and bought the Maestro 7 product from Unison. That product evolved into TWS 8.2 / 8.3 / 8.5.1 / 8.6.

Initial setup was complex; it involved automating huge complex business process models for (nowadays legacy) order, contract and invoice management applications, although limited to mainly workloads for the HP-UX operation system. Later on, the models became less complex, but the agent landscape became more heterogeneous (Solaris, RedHat Linux, SUSE Linux, HP-UX, Win 2003, Win 2008, SAP R/3, SAP BI).

Implementation has been a joint effort with IBM Tivoli L3 support, PostNL IT RHEL and Windows administrators and a dedicated WLA coördinator / TWS administartor

No idea how to calculate the ROI in Euro's or Dollar's, but for sure it was worth the investment.

Make your TWS infrastructure auditable by IBM (or to its partners) to be able to benefit from sub-capacity PVU licencing when your networks use a considerably amount of virtualization technology, but above all, when possible, move to cloud-provided TWS management services, to benefit from more modern ‘pay-per-use’ licensing models.

We considered starting with Redwood Cronacle for WLA on SAP, but decided to stay with TWS because of the better integration between SAP and non-SAP workloads, and the minimal amount of effort we had to put into education.

Just contact IBM sales and make the first implementation of the product at your site a joint effort between IBM Tivoli L2 personnel from Rome and your WLA administrators.

The most valuable feature is the job scheduler.

I would like to be able to access the return value or result from one job, in the following job. This is a feature that other solutions have and is very helpful. I had to create my own workaround for it because the capability is important for me.

It would be helpful to have a mobile app that could be used to follow the job schedule. Most IT applications now have mobile app support.

I worked on IBM Workload Automation for close to 10 years, although I have recently changed companies and no longer use it.

The stability is good. 

We did not need to scale this solution. We had close to ten people who used it.

Technical support from IBM is very good. I have worked with different vendors and I know that it is sometimes difficult to get a good level of support, but there were people supporting this solution who were very good and always helped.

The implementation was done by external consultants.

I would rate this solution an eight out of ten.

• Event-based scheduling
• Support for other applications via connector, such as Oracle, SAP and data warehousing software

If you want to run any Oracle, SAP or data warehousing job, you don't need to install a TWS client on the target server. You just have to install a connector on one of your TWS systems and provide the required parameters such as system credentials in an encrypted format.

It supports almost all of the different groups in the organisation for automating their tasks, such as finance and accounting, transport department, procurement department and several others as well.

Whenever we find any bug/vulnerability, we immediately inform IBM and they provide us a fix, so I can say it is being continuously improved.

A few things are missing but I can manage without them, such as a cross-reference report. For example, if you want to find particular job script, it is very difficult to do so from the TWS command line or from the TDWC (Tivoli Dynamic Workload Scheduler) tool. On the other hand, there is a Germany-based company Horizont, which has a very nice tool, TWS/WebAdmin, which provides such a facility.

The second thing is that there is no built-in facility to run SFTP/FTP file transfer jobs. You have to create your own SFTP/FTP scripts to run these types of jobs.

An internal change management tool is also not available in IBM TWS, but other scheduling software do offer one.

These are a few flaws in TWS, but operation-wise and handling-wise, it is a superb tool as compared to other batch job scheduling tools.

One thing is true: You will not get all of the facilities in one tool, so I recommend you go with TWS for smooth automation delivery.

I have used this solution for the last seven years.

We never encountered any stability issues. If your TWS administrator is very good with TWS, then he/she should always keep your TWS landscape stable.

We never encountered any issues with scalability.

We used BMC Control-M first, but later we changed to TWS due to its efficiency, accuracy and scalability.

If you have strong knowledge in TWS, then initial setup is just like bread and butter for you.

I heard it is very expensive compared to other job scheduling tools.

Go ahead and use the TWS solution. It is really a very good product.

We use it at the Bank for the portfolio management system. In 2001 and 2003 I created all these jobs for the portfolio management system for collecting data from the backend system.

From 2015 I use it as the tool that collects the resources within the IT system that reports the use of different resources, file system resources, storage resources, and so on.

The most important feature is the creation of folders. It's a really great feature because you can organize the process with naming conventions. 

Other solutions like Control-M are better than this solution. IBM should have better integration with the cloud. 

It should support other schedulers that aren't IBM products. 

We have been using IBM Workload Automation since 2015. 

It's very, very stable. The support is quite good. I was totally astonished about the life cycle of this IBM product because a lot of colleagues told me that Control-M is the best scheduler but I have found that it's not true. IBM does a really good job with its product.

It's better than the solution that Control-M offers. Within Deutsche bank they are using Control-M for the investment banking and they are using different timezones, for every console you need a separate installation with Control-M. With IBM, it's not necessary. 

IBM uses also has different plan options, from monthly plans to weekly plans. This is a problem with Control-M. It makes maintenance very, very costly because you have to configure different environments for different timezones.

The scheduler is a little bit complex. They have different concepts, they use event mechanisms for their workforce. 

With Control-M you have a lot of server components. In the meantime, you can install them with Kubernetes as well which is a big improvement.

I would definitely recommend this product. It's an interesting option in terms of schedulers. 

I would rate it an eight out of ten.

• Alerting on ABENDS: When a job abnormally ends (ABENDS), the solution notifies us when the job did not complete successfully. This is a nice feature for job streams that require successful completion of one job before it moves on to another. It’s still “reactionary” in nature, but allows us to run a job stream again. This occurs, in some cases, before the end users of our data services know there’s a problem.
• When changing a cycling ID, we only have to change it once.

• Moving jobs from Dev to Prod only takes a text file script, which is straightforward.

It has over complicated things.

• It is complex to use
• Lacks scalability
• It is difficult to set up jobs to run
• We need to be able to elevate privileges like Task Scheduler

With Windows Server 2012, there is a setting for User Access Control. UAC is a security feature that prompts the user, and even administrators, when running a job that requires a higher elevation because of its interaction with the Operating System.

When you’re trying to automate tasks and run them in the background, this prompt still comes up and asks if it’s OK to raise the privilege level. That hangs the job.

When running the solution, there’s no way to get around that prompt. Any job that requires that elevated privilege sits and waits for the user to answer it.

With Windows Task Scheduler, there is a check box in the Scheduled Task that says “Run with Highest Privilege”. If that’s checked, then it automatically bypasses the UAC prompt, and completes successfully.

The only other way to get around that is to lower the UAC restriction on the server, making it more vulnerable to malicious code.

I have been using this solution since November, 2010.

We have not have issues with stability.

We had to write a scripted solution to be able to change jobs behind the scenes. The solution runs with a specific command fed into it. We fed it this script: “CSCRIPT

Our company has our own internal support.

We used Windows Task Scheduler. We switched because the new SSIS servers were going to be managed servers.

The setup was very complex. Our environment is very large. Setting up Workload Automation on our set of servers required the following:

• Setting up the agent
• Setting up the notification lists
• Filling out various forms for job stream scheduling. (It just goes on from there.)

This was the new corporate standard and we were not given a choice.

Look into ALL options. Verify that you can run your application servers with UAC turned to "do not prompt when using an admin account".

TWS has two command line interfaces to help manage items in the active plan and database. This allows you to create custom scripts to make mass updates to job definitions in the plan\database. You can even schedule these custom scripts to make the updates at a later time (e.g., holding a set of jobs prior to a planned outage and releasing after). It also has an event processor that lets you define custom rules based on certain events (e.g., when a job fails, when a job hasn’t finished within a certain time or within 30 minutes of its predecessor, when a file is available, when a job is stuck, etc.).

Aside from basic scheduling features and Windows/Unix job types, this product gives us a lot of options to integrate with other systems:

• Custom plug-ins (SAP, DataStage, PeopleSoft, Cognos)
• Database & integration job types (Oracle, MS SQL, Web Services, J2EE)
• File transfer and remote command job types

It is a great way to centralize your enterprise workload to provide better visibility to support groups and better cohesion between applications. Mobile app features are also available to empower users to monitor\manage their workload on the go.

The web UI is a bit cumbersome in terms of navigation, although it seems some improvements were made on the most recent release, v9.3. Also, scheduling is driven based off of a Symphony file, which is a flat file that outlines scheduled work for the day and its progress. This file is prone to corruption during networking/hardware issues. TWS also has a daily refresh process that refreshes the Symphony file each day. Any permanent changes to the TWS network are dependent on this process.

Time zone management can also use some improvements. We have to schedule creatively to account for the hard-coded plan start/end times with other time zones. Ideally, it would be a seamless effort.

I have used it for four years.

None; overall the product is very stable, but can give you odd results if the environment is unstable (hardware\network issues). Once the environment issues are resolved and job scheduling has resumed, you might still experience strange behavior with dependency resolution and cyclic job functionality, which typically clears up with the next daily refresh.

None; we’ve had no issues with scalability other than time zone management, which requires some creative scheduling scenarios to account for the hard-coded plan start/end time and time differences between master server and remote agents.

7/10; technical support is virtually impossible to get via direct phone call. You have to rely on trouble tickets and wait for a return phone call, which can take several hours depending on what time of day you are calling. You will need to keep constant follow-up on your ticket, as well, because it can slip through the cracks and go unnoticed for days. However, once engaged, especially on a conference call, the team is knowledgeable, courteous and very helpful.

Initial setup was somewhat complex, although it might be because we installed on Linux. Windows might have been a bit easier using the installation wizards.

It can get costly, as expected; IBM uses a Processor Value Unit (PVU) licensing model, which can be cumbersome to maintain.

I’ve used BMC Control-M in the past, at another company.

This product is a great product overall, but can behave strangely if any environmental issues occur (hardware/network). A lot of the issues we face, however, are supposed to be resolved with the later version. I’ve seen some demos of the latest release and it seems like large improvements have been made in terms of functionality and user experience.

The whole product is valuable because it is a tool for batch automation.

There should be more custom documentation, specifically around Java APIs. There should also be more training.

In terms of features, we are currently using only 50% of its features. We don't use all features that are available, but there is always room for improvement in all of the tools. 

I have been using this solution for 15 years.

It is very stable.

It is extremely scalable. We run more than 80,000 jobs a day across multiple platforms.

Their technical support is fine. We have a contract with them to assist us.

It is simple to install.

It is about one-third of the cost of a controller.

I would rate IBM Workload Automation a seven out of ten.