Smart Cloud Automation: Are you Scaling In and Saving Costs?

At a session for SAPInsider Events, Brenton O’Callaghan, Chief Customer Officer, and Tyler Constable, Director of Sales Engineering at Avantra discussed the future of Smart Cloud Automation and how businesses can save costs by scaling in.

From using Smart Cloud Automation to save costs, to companies moving towards an automation-first approach, Brenton and Tyler spoke in-depth about how organizations can drive value from a hybrid cloud investment. Read the highlights of the conversation below. 

What is Smart Cloud Automation?

Brenton O’Callaghan: Based on your experience working with customers during the sales process, what is smart cloud automation and why is this a conversation driving topic for customers?

Tyler Constable: Traditionally, the word ‘scale’ and SAP have never really come together. 

As opposed to having to purchase the necessary hardware earlier, the availability of public clouds now has made scaling possible and customers are just beginning to realize this. The benefit is that customers can now use resources as per their need. They can scale up, automate the scaling and even scale it all back. 

Companies, large or small, also have the option to cut down on costs if, for example, they do not need some or all non-production systems running throughout the week.

While they would still need to pay for other measurements, over the course of a year the amount of money saved with smart cloud automation is quite significant. 

It also greatly helps developers. They do not have to depend on the Basis team anymore to shut down systems and bring it back up during weekends, they can even have partial or whole systems running according to their need.

Brenton O’Callaghan: Going in the opposite direction, in a scenario where employees need to automatically spin up an additional application server; that would add massive value to their business, but how does cost factor into this? 

Tyler Constable: In that scenario, even though there is a cost increase, companies are saving on a lot of other resources like time - by not having to go through multiple levels of approvals and waiting to procure the necessary hardware. 

With a platform like ServiceNow managing digital workflows, you can have an additional application server in the shortest time possible, with minimal human intervention instead of waiting for months on end to get approvals and then set everything up. You only need to maybe just click a button on the phone and the rest of the process is automated for you.

Scaling: In, Out, Up, Down

Brenton O’Callaghan: For clients who may not be familiar with SAP operations in depth, how would you define terms like scale in, scale out, scale up and scale down in the context of automation? 

Tyler Constable: From the point of view of a single server, scaling up would mean adding more resources to that server while scaling down would mean to remove resources from it. Public clouds can scale up and down automatically and when needed.

Scale out, on the other hand, is not supported by public clouds and this is when you add additional application servers that already exist in the system or are pre-built. If you have 3 out of 4 application servers in use and need an additional one, you can use the idle server with the help of a third-party solution such as Avantra. Whether it is Google, AWS, Azure or Alibaba, it is the server itself that scales with the SAP specific application that sits on top of that server. 

This is highly cost effective too. It is much better to have a server, switched off and waiting to be powered on, available in the system, over having it perpetually switched on and just waiting to be connected to. With public clouds, you are paying per resource or per minute. Whereas, with a server, you are just paying for the disk space and not for the other resources.

The Journey to the Cloud

Brenton O’Callaghan: The transition to the cloud is a rather long journey that spans years, and it is not an easy one. How does a hybrid scenario, where having both on-premise and cloud solutions during this transition, work for customers?

Tyler Constable: Every single one of the 4 major cloud providers understands that the journey to the cloud is not one that will happen overnight. This is a process that takes 1-2 years. In such a scenario, going the hybrid way is the only way out.

In this case, being able to see what is on premise and what is on cloud, being able to start up and shut down all of that from a single management system is important. Not just from the public cloud perspective, but also on premise as well to have a comprehensive view and manage both systems, at all times.

Brenton O’Callaghan: Building on that answer, with start-stop automation just being the tip of the iceberg, what other scenarios or types of automation can customers expect? 

Tyler Constable: I think the one that people are most interested in is OS patching and fully automating that process. This used to be a very manual process, but now you could schedule that without even having to sit up and monitor it. 

Brenton O’Callaghan: But when it comes to hybrid management, how important is it to have integrations in place to notify someone when there is an issue in the system?

Tyler Constable: Having a third-party integration service like PagerDuty or ServiceNow will be crucial to start and stop the server, and to also raise issues and notify the stakeholders. 

This way you know why a process or even a batch job failed. You could pinpoint the reason why it failed, saving employees’ time. Companies should look for a third-party solution that can look at the system thoroughly and then take the appropriate action of automating or notifying.

Brenton O’Callaghan: To add on to that point, lowering costs are synonymous with decreasing the employee count for some companies. However, it should be about freeing up human capital to focus on activities that ultimately add value to their business. With automation, this is possible as employees do not have to tend to the same problem multiple times when they can just automate the solution.

This process not only lowers costs, it also saves time and prevents saturation too. By taking advantage of tools like Avantra for new automations, you will have more time at hand to innovate. 

Coming back to companies in the hybrid scenario during the transition, assume a client’s ECC system is sitting on premise, and there is a requirement for a couple of additional application servers in the cloud. What are the different routes they can take to get value from automating in the cloud during this period? 

Tyler Constable: As long as the networking is in place, you would still have an application server up in the cloud. 

Even if the primary application server is still on premise, the system would be able to go out onto the public cloud and bring up the sleeping application server. It's not just that your hybrid environment is just 'BW up in the cloud and ECC is on prem', but it's the actual system itself that becomes hybrid.

Brenton O’Callaghan: How important is the quality of the pipe between your on-premise system and your cloud infrastructure, then? Should you make sure your data center from the cloud provider is nearby?

Tyler Constable: It is important to start to architect what the public cloud scenario will be even before you start the hybrid scenario. As you architect that solution, you need to think about using minimal resources and also have a workflow in place. You can select which public data centers you want based on its proximity to the cloud provider. 

This transition is not as easy as just ‘forklifting’ systems; rather it must be a thought-out process. At every step when moving something to the cloud, ask yourself what value you might get from it and work towards that. Do not wait till the end to realize the benefit as you will be missing out on so much at every step in between.

Brenton O’Callaghan: You mentioned that planned maintenance and automation can go side-by-side. Is completely automating your maintenance windows possible with OS patching, for example?

Tyler Constable: Absolutely. Not just OS patching, you can get creative with what you want to automate down the line. It could be a test within SAP or a talk between bank interfaces. They can all be fully automated and done in maintenance windows. All the traditionally manual steps that you have done over the years can be automated now. 

This drives consistency and enables traceability as well. Since everything is automated, you no longer have to depend on one particular employee to do the job. You also have access to the complete log of information and reports thereafter, to help you understand and monitor the situation.

On the other hand, if you would like to have more control, you can enable systems to start up and shut down manually at the click of a button, without scheduling or automating the whole process too. Also, this helps non-technical managers gain a holistic view of their landscape.

Brenton O’Callaghan: Do you have any last thoughts to add before we close the session?

Tyler Constable: While we majorly focused on the application of automation in an SAP environment in this session, the point to note here is that whatever we discussed earlier today similarly applies to non-SAP applications too. All these applications and third-party solutions can help save your cloud costs, especially by not running all the time.

Photo by Micheile Henderson on Unsplash

Further reading: 

Cloud infrastructure automation