AWS Core Services & Global Infrastructure for SAP

You need an SAP app server running in a second AZ for HA. Your VPC already exists. What do you actually have to create, and what is the common misconception about VPC/subnet scope?

An SAP HANA volume is running low on IOPS in production. A colleague plans to stop the instance to grow the volume and bump IOPS. Why is the downtime unnecessary, and what's the one capability that still carries a volume-type restriction?

For SAP DR planning, what does the ~100 km AZ separation actually buy you, and why is spreading across AZs NOT a substitute for a second Region?

Your SAP backup automation fails creating an S3 bucket with 'BucketAlreadyExists' even though you've never used that name. What scope is the name unique in, and what hard object limit should backup tooling respect?

An SAP-adjacent workload must stay physically in your own datacenter for a data-residency/security mandate. Which AWS extension fits, and how is it different from a Local Zone or Wavelength Zone?

Which EC2 placement group do you choose for an SAP HANA scale-out cluster, why is the fault-tolerant option the wrong instinct here, and do placement groups improve your EC2 SLA?

When you move an SAP-adjacent component from IaaS toward PaaS/SaaS, what exactly do you trade, and why isn't 'more managed = always better' the right rule?

AWS quotes 99.5% instance-level and 99.99% region-level EC2 SLAs. What is the subtle catch in how the region-level number is measured, and what does that imply for an SAP single-instance app server?

A stakeholder calls 'S/4HANA on AWS' a SaaS deployment. Why is that wrong, what is your only responsibility under genuine SaaS, and which SAP products actually are SaaS?

What is the minimum AWS Support plan to be eligible for SAP support, and why is the Developer plan a trap even for a non-production SAP sandbox?

An architect proposes running the SAP application tier in containers on AWS for portability. Why is that not viable, and which less-common compute forms ARE supported?

You need certified, low-latency shared storage that can serve SAP HANA as both file and block. Why pick FSx for NetApp ONTAP over plain EFS, and what cross-region capability comes with it?

Beyond billing, what two other boundaries does an AWS account enforce, and why does that shape multi-account SAP landing-zone design?

An SAP system already has HA via a Pacemaker cluster. Why must you DISABLE EC2 autorecovery, and which two other instance settings matter for production SAP?

For a typical SAP volume, why is gp3 the default rather than io2, and which volume types are disqualified from being a boot volume?

You launch an SAP EC2 instance from a 'bring your own image' AMI. What licensing obligation does that carry that a Marketplace SUSE/Red Hat AMI does not?

For an SAP DR target that sits powered-off until needed, why does a Reserved Instance fail to guarantee it'll launch, and what actually does?

Among Elastic IP, ENI, enhanced networking, and ENA Express, which gives you a detach/re-attach virtual NIC for HA flexibility, and which two are about raw network performance?

On a new SAP-on-EC2 system, transactions OS06/OS07 show missing or wrong OS/storage/network data. What component is almost certainly not installed, and why does it also block support?

An SAP-on-Windows landscape needs shared storage for a SQL Server HA configuration. Why is FSx for Windows File Server the fit rather than EFS, and what specific feature makes it work for SQL HA?

Why can't you attach an existing EBS volume to an EC2 instance in a different AZ, and what are EBS's volume-level vs region-level SLAs?

EFS is great for shared SAP directories like /sapmnt on Linux. What hard limitation stops you from using it landscape-wide, and what's the multi-AZ payoff when you can use it?

For SAP backups, what decides between S3 Standard, an IA class, and a Glacier tier — given all of them share the same durability?

Why are SAP-certified EC2 instances built on the AWS Nitro system specifically — what does Nitro remove that matters for SAP, beyond being 'newer hardware'?

Under the IaaS shared-responsibility split for SAP, where exactly does the line fall — and what does that mean for guest-OS patching?

Across EC2 purchase options, which one must you NOT use for SAP and why, and what's the trap in assuming a Reserved Instance guarantees you capacity?

Decode r6i.8xlarge and c7gn into family/generation/processor/size — and name the two reading errors people make on the size and the 'g'.

An SAP component needs raw TCP load balancing, not HTTP. Which ELB type fits by OSI layer, which one would you wrongly reach for, and what health rule do all ELBs follow?

You want SAP traffic to reach an AWS service (e.g. S3) privately without the public internet, and separately to connect many VPCs through one hub. Which option is which — and what's the classic mix-up?

Availability Zone Independence covers physical isolation — but how does AWS extend the SAME principle to its own software/service updates, and why does that matter for your multi-AZ SAP design?

A team wants to move their SAP database to Amazon RDS to offload DB administration. Why does that fail for most SAP databases, and what's the narrow exception that actually works?

What do the key EC2 instance-family letters mean: C, M, R, X, U, I, T?

You find a blog claiming an EC2 instance type is SAP-certified, but it conflicts with a SAP Note. Which Note wins, what's the other foundational Note, and what scope do these Notes NOT cover?

Beyond gateways and peering, two VPC constructs come up in SAP networking: one decides where traffic goes, the other records where it went. Which is which?

After picking a region for a new SAP landscape you notice AWS already created a VPC there. Why shouldn't you build SAP into it, and what is it for?

You're designing SAP DR and read that AWS gives an SLA for the data plane but not the control plane. What's the difference, and why does it change how you plan a region-failover?

AWS offers 15+ managed database engines, but they fall into distinct categories. Match the category to its AWS offering — and note what this means for a certified SAP database.

Why does an EC2 instance type need to be 'certified' for SAP at all — what role does AWS play in SAP's eyes, and what does certification unlock?

You can read the family letter in r6i, but in a name like c6gd or m6in, what do the processor letter and the trailing capability letters actually tell you?

Your SAP HA design assumes one Amazon EFS filesystem can be mounted by app servers in two different VPCs, and also replicated for DR. Which of those is a hard limitation, and what does EFS actually offer for DR?

The Amazon FSx family includes four file systems. Which two matter for SAP on AWS, and why can you ignore the other two?

AWS classifies services as zonal, regional, or global for fault isolation. Place EC2, SQS/DynamoDB, and IAM into the right bucket — and how can you often tell zonal from regional just by starting to create one?

How does the book quantify AWS's global reach, and what precisely is an AWS 'region' as a building block of that footprint?

When AWS decides where to place a new region, what factors drive site selection, and what resiliency property is that placement engineered to deliver?

Which operating systems does AWS support for SAP workloads, and why does picking outside this list forfeit support?

An SAP HANA node isn't getting the network throughput you expected, and a colleague suggests flipping a network setting. On EC2, what fundamentally determines an instance's network bandwidth?

During an AWS service interruption you find you cannot launch new EC2 instances, yet your already-running SAP systems keep serving traffic fine. Which 'plane' is impaired, and why does AWS split them this way?

A colleague assumes any SAP product can simply run on AWS. AWS actually publishes a defined list of supported SAP products — what kinds of products are on it, and where do you confirm exact versions?

Some AWS services start with 'Amazon' (Amazon EC2) and others with 'AWS' (AWS Outposts). Is there an official rule behind the prefix, and what does the community heuristic claim?

Your SAP application server is undersized. On AWS, what does resizing the instance actually require of the application, and which performance dimension besides CPU/RAM improves when you move to a bigger size?

SAP sizing calls for 256 GB of RAM, and the EC2 spec sheet lists the instance at 256 GiB. Are those the same number, and which way does the discrepancy fall?

You're provisioning a large SAP landscape and EC2 launches start failing partway through, even though the instance type is available. What account-level constraint bites here, and what's the remedy?

Map AWS's three storage categories to their SAP roles: which services are block, which are file, which are object — and what SAP use case does each serve?

AWS says you rarely need RAID on EBS because you can tune performance dynamically. So when does RAID 0 striping still make sense for an SAP volume?

EFS offers Standard, Standard-IA, One Zone, and One Zone-IA. Since all four share the same durability, what actually separates them when you're choosing shared storage for an SAP HA design?

Two SAP teams share storage on FSx for NetApp ONTAP: one wants to minimize storage spend, the other wants a 99.99% availability commitment. Which ONTAP features and deployment choice serve each?

You're sizing FSx for Windows File Server for an SAP-on-Windows landscape. What is its max filesystem size, throughput, and the multi-AZ vs single-AZ SLA — and when does that ceiling push you elsewhere?

Within an SAP-on-AWS landscape, which service registers domain names, routes traffic by DNS, and health-checks resources — and what makes it more than a plain name server?

'AWS already gives me HA at the data-center level, so cross-AZ connectivity can't break.' How far is that true, and what redundancy actually backs cross-AZ traffic within a region?

You want one SAProuter with a fixed Elastic IP to serve SAP systems spanning two AWS regions. Why doesn't that work, and what does it force on your SAProuter design?

An SAP automation script needs the instance's own AMI ID, IAM role credentials, and IP without hard-coding them. Which EC2 service provides that, and why prefer IMDSv2?