Cloud resiliency is driving cloud adoption.
SAP has made no secret of the fact that they plan to be a cloud-first company. Moving to cloud has many advantages for businesses, including flexibility and scalability. It also represents an opportunity to reinvent business models, access Big Data, and adapt processes in real-time. But for many businesses, resilience is the driving factor behind the adoption of the cloud.
Businesses simply can’t afford any application downtime. Expectations demand an ‘always-on’ experience. There is clear need to be able to recover and continue operations without systems going down, losing functionality, or creating blank spaces in their real-time processing and data analysis.
Leveraging the cloud for more resilient infrastructure
With the end of support coming for ECC, migration to S/4HANA is expected to make moving to the cloud easier – and increase its uptake among SAP clients. Speaking in SAP’s Q2 earnings report, SAP CEO, Christian Klein explained their investment in resilient cloud infrastructure.
‘We have decided to speed up the modernization of our cloud delivery to enable a more resilient and scalable cloud infrastructure. This will require additional investments in the next 2 years.’SAP CEO, Christian Klein
While the Cloud poses new challenges for companies in terms of business continuity, the cloud also brings new options for the creation of Disaster Recovery Plans that allow a company to restart its IT after an outage or failure that shut down its systems or made them inaccessible.
However, to leverage the Cloud in this quest for resiliency, it is necessary to implement practices adapted to the characteristics of these environments, and accept a few key principles, such as the most accurate sizing of machines and the systematic use of automation, with infrastructure as code which allows you to control IT systems via APIs.
DRP-as-Code solution for SAP environments
There are many Cloud-specific features that can make your Disaster Recovery process more secure. These strengths have been heavily used by PASàPAS to develop DRP-as-Code (a new generation Disaster Recovery Plan) as part of its SAP Cloud Managed Services offer.
“As soon as the Cloud became an option for SAP systems, we identified it as a great way to address business continuity issues,” say, Brian Passante, Cloud & Innovation Manager at PASàPAS.
The solution combines the best of Cloud security, leveraging the secure foundations offered by hyperscalers, and additional services to secure all the infrastructure layers and software components. So much so that today, DRP on the Cloud has become a native component of PASàPAS’ SAP Cloud Managed Services offer, based on a multi-cloud backup solution adapted to SAP software environments.
Minimising downtime and data loss with a Disaster Recovery Plan
Downtime and data loss are two key parameters of any Disaster Recovery Plan. As Jérôme Mollier-Pierret, director of SAP outsourcing at PASàPAS, explains:
“Most of the systems we deal with are business-critical systems, which we can hardly ever stop and for which tolerable data loss and downtime are minimal”.
The Recovery Point Objective (RPO) defines the quantity of data that it is acceptable to lose, therefore the freshness of the backups that must be kept permanently. The Recovery Time Objective (RTO) then represents the maximum tolerable duration of a service interruption.
Next to consider in a potential DRP is that SAP systems typically handle large amounts of data (commonly several TB) and HANA architectures require vendor-certified machines with large amounts of RAM.
“This is another specificity to take into account” adds Brian Passante: “only about ten percent of the servers offered by Hyperscalers meet these criteria.”
“The number of interfaces is also essential in the construction of the Disaster Recovery Plan,” continues Pesante, “because, more than the restart of back-up machines, the real objective of a DRP is to make applications accessible to users, including third-party applications”.
Moreover, when several SAP applications are interconnected, it will be necessary to restore them simultaneously, to the second, and ensure the reconstruction / continuity of the associated application flows in complete safety.
Restoring a full SAP Production system in 4 hours
These are all elements that PASàPAS has integrated into its DRP solution. A solution that is now available and successfully tested by a client which has used PASàPAS to outsource their SAP systems on the Cloud.
This customer, who specializes in the recycling of electrical and electronic equipment, was able to test the effectiveness of its DRP in the Cloud. This consists of a complete SAP Production system, made up of 8 applications and 17 TB of data, in less than 4 hours.
“In simple terms, our solution makes it possible to rebuild the entire customer’s infrastructure in a completely automated way and to restore the latest SAP system backups, with an RPO of less than 30 minutes“.
Trigger DRP tests without excessive costs
And, unlike a traditional DRP, this model does not involve any dormant infrastructure. Consequently, triggering a test of this plan costs less than 0.5% of the annual infrastructure bill.
“This means that customers understand that they can test their DRP more often, replaying it regularly, even automatically.”
In a world where ransomware attacks have become more common, this is a plus for companies audited on the resiliency and security of information systems.
The end goal of the model is a completely resilient infrastructure, self-repairing, regardless of the component affected by a malfunction. “Our team is already operating with this mindset,” he says, “even though SAP is not Netflix!”. Resiliency is so crucial to Netflix, they will run disruptive software (such as “Chaos Monkey”) on their production information system.
DRP-as-Code works natively with the main Clouds and is ready for simultaneous events
Now deployed to clients managed by PASàPAS, DRP-as-Code covers all types of SAP environments (especially ECC 6 and S/4HANA). It is able to leverage all the new capabilities offered by the Cloud.
So far, the PASàPAS DRP solution has never been triggered for a “real” disaster. However, PASàPAS does anticipate simultaneous events, potentially affecting several companies at the same time. “This is the reason why we chose to automate as much as possible. In order to be able to absorb such an event”, explains the Cloud & Innovation manager of PASàPAS.
In addition, in this perspective, the hyperscalers’ infrastructures do offer a certain number of guarantees, unlike traditional infrastructures, where the simultaneous restoration of large volumes of data very often results in a performance degradation, putting the RTO at risk.
“One of our next steps is to extend the scope of the post-restart phase to cover specific elements related to our customers’ non-SAP systems” notes Jérôme Mollier-Pierret.
To reduce the dependency on a particular infrastructure, backup systems are hosted on a different Cloud. Our DRP-as-Code works natively with the main Clouds on the market (Amazon Web Services, Microsoft Azure and Google Cloud). But switching to a multi-cloud DRP raises some additional complexities, especially for networking.
If you would like to learn more about cloud resilience and DRP-as-code, click here to speak with our team.