Welcome to this quarter’s edition of our SAP on AWS Innovation Newsletter. Lemongrass’s regular update details the new features, capabilities, and market updates that drive innovations and optimizations into SAP-Centric workloads on AWS.
I‘m sure you are also looking at the calendar, wondering how we arrived at Q2 2021 without the aid of a DeLorean uber ride from early 2020. Remote working hasn’t hindered the innovative home office keyboard warriors releasing some fantastic new platform innovations into our SAP on AWS universe. We can be sure that as the world starts to unlock, we’ll emerge from home hibernation into a more technically optimized planet outside our front doors. Here are a few of our favorite announcements that have made an impact on our customers over the last quarter:
- A big update for our SAP-Centric world was AWS EFS which has been subject to a significant update with the introduction of the EFS one zone storage classes, which can reduce the storage class cost by up to a whopping 47% compared to standard EFS classes. The little pay-off is that the durability is reduced from 11x9s to 3x9s, as it’s a single AZ vs. regional service which is sufficient for most non-prod workloads. However, it is worth noting that One Zone EFS systems are by default backed up with AWS backup, giving a default recovery model (which you are able to switch off)
- EFS recently had a large upgrade: 3x higher read throughputs are now possible, which follows in the footsteps of the GP3 flex models launched in December 2020. For example, activating the burst mode file volumes can achieve 300mb/s (300mbs per TiB, to be exact). It can be further boosted if the volume is configured with 1GB/s provisioned throughput, which offers the same acceleration multiplier taking it up to 3GB/s for reading operations. Best of all is the throughput uplift. This has been automatically enabled for all file systems in all regions at no extra cost – Lemongrass look forward to sharing this with our customers in our monthly optimization sessions
- AWS continues to invest in Automation in the SAP space. Last month the SAP Launch Wizard gained the capability to deploy both SLES & RHEL distributed clusters for the SAP application layer. Before this release (although it distributed installs), it required the manual config at the OS level. It now has the functionality to fully support clustered deployments for single & multi-node deployment models across AWS Availability Zones. The new capability means it is now an extremely powerful tool for production-grade greenfield deployments
- AWS has been fast-developing “Low code” UI’s to enable non-developers to access and leverage AWS tooling that historically was at the realm of coders. Lambda recent makeover had been in beta for the past few months and is now the new permanent UI since the end of March 2021. The UI is far more “Drag & connect” in nature, thanks to a new configuration tab that can reduce the screen space required to build new Lambda functions. Lemongrass has found this very beneficial in enabling some of our technical architecture community to create advanced automation functions easily
- While on the topic of Lambda, it is now possible to programmatically run four of the primary Trusted Advisor checks across accounts using a simple Lambda trigger. The update covers the check for high error rates (+10% day growth), Excessive timeouts (+10% over seven days), checking for depreciated (or about to be retired) AWS runtimes, and finally, any Lambda functions that are running in a single AZ that might be sensitive to service interruption. For example, Lemongrass used this recently to check for any Python code running on version 2.7 in just two drop and drag queries across our entire install base
- As Lemongrass has seen many times, storage costs are often one of the areas that, if not managed correctly, can be a costly mistake. Recently, we have seen a few key updates that will help this problem. Firstly, a new version of BackInt (1.03) uses this to write backups directly into S3, avoiding expensive intermediary storage types. This new version enables backup object tagging (so we can add meaningful names), scheduled log rotation giving us finer-grained control on log file management, and automatic signature verification allowing the ability to provide the S3 bucket location to the signature file for enhanced security. Secondly, AWS Glacier has just released a 40% discount on pricing for PUT & lifecycle, which will further help data aging into deeper cost-optimized models
- A beneficial new FinOps update has been announced with the native AWS Cost Anomaly Detection console. This update further supports cost monitors and alert subscriptions via AWS CloudFormation templates, essentially meaning it is possible to build cost Anomaly Detection directly via YAML/JSON commands. The powerful enhancement enables cost spike alarms to be built directly into the provisioning template, which is read and integrated into the Machine Learning runtime which sits inside the Cost Dashboard. It provides cost management alarms built directly into SAP automation executions merging provisioning & FinOps in just one click
- For any customer with significant concerns or equally large traffic profiles, AWS Direct Connect is a major beneficial update. It now offers native 100GBPs dedicated connection options and fixes the problem that’s driven some customers who need extreme connectivity to parallel wire in multiple 10Gbps connections in a linked aggregation group. Organizations that benefit from this update are those who fire over large amounts of time-critical data (e.g., industries including Utilities / Media & broadcasting / Financial services, and many others). The major change of providing more extreme comms lines is only available in certain regions, and I’m sure this will be rolled into all 24 regions through the year. The pricing is slightly higher, but it’s still competitive when you consider the size of the connection port you’re dealing with; a 100G port which will be priced around $22.50 / hour
- EBS CloudWatch metrics now have finer grain resolution and are down from 5 minutes to 1 minute (previously only available on Io2/1 volumes before). It is a significant change for Lemongrass, as this was one of the reasons why we couldn’t ever rely on CloudWatch exclusively for monitoring. Losing 5 minutes in a 15 minute RTO was just not feasible. The new resolution across all EBS volume types means we are now moving back to CloudWatch for critical volume monitoring and leveraging the box CloudWatch alarms into SNS. If you have CloudWatch active in your accounts, congratulations, you’re now 5x more responsive!
- Upgrading EBS metrics without improving SNS metrics would move the reporting resolution problem downstream, so AWS has increased the SNS publication metrics down to 60 seconds. Again, it’s a no-click service update, so if you have any SNS subscriptions setup –check out the 1-minute notification update from the original 5-minute baseline
- SAP has announced a Content Server for EFS with MaxDB. It is an easy update to any customers running CS7.x and represents an excellent optimization step vs. running on a traditional Database instance
- Finally, the R5b EC2 family accomplished the SAP-certified list and has been updated on the AWS certification site. While these systems appear similar to the base R5 family, they do pack some more punch at the EBS level using the new 60GBPs network for the larger systems, and the 10GBPs for the small systems. It also gives access to the boosted 260k I/O to better manage improved storage-based workloads than the base R5’s, despite having near-identical SAPs ratings it offers a great performance driver for intel HANA workloads
As always, there are many more relevant updates than we can compress into one newsletter, and it’s only fitting we give an innovation appreciation salute to the many biochemists & scientists who achieved the impossible by cooking up extremely complex Vector & RNA vaccinations in record time that have given us a roadmap to normality.
Keep safe, but do keep on innovating…
Ben
