Amazon EC2: Modify CPU options for Windows license-included workloads

How CIFS made Windows file sharing consistent before the cloud era In the late 1980s, Microsoft built CIFS, the Common Internet File System. While Sun’s NFSv2 chased stateless speed, CIFS chose stateful consistency. That decision still shapes how Windows shares files today. → CIFS servers remember every session, user, and open file. → Each connection carries context, UID for user, TID for share, FID for file. → Operations can be batched (ANDX) to save network time. When you open a file, CIFS enforces share modes and mandatory locks. No guesswork. No conflicts. It’s strict by design. To stay fast, CIFS added opportunistic locks, oplocks. → Exclusive locks let one client cache writes. → Shared locks allow read caching. If another client wants access, the server revokes the oplock and forces a flush. If the TCP link dies, CIFS assumes failure and stays consistent. NFS waits. CIFS acts. The tradeoff was simple: → Always consistent. → Usually fast. That mindset still powers SMB3, the protocol behind modern Windows and Azure file sharing.

Amazon GameLift Servers adds telemetry metrics to all server SDKs and game engine plugins - Today, Amazon GameLift Servers launched the addition of built-in telemetry metrics across all server SDKs and game engine plugins. Built on OpenTelemetry, an open source framework, Amazon GameLift Servers telemetry metrics enable game developers to generate, collect, and export critical… https://lnkd.in/ejvdQrEN

Infracodebase now supports the official GitHub and HashiCorp #Terraform #MCP servers. This means that you can use your companies private modules to build your app, landing zone, and other infrastructure. This has many benefits including: → Less friction to adoption of your shared services → Faster discoverability and integration → Less focus on supporting the basics so you can invest in the hard stuff → Reuse of your secure-by-design infra system at scale What tools would you like to see us onboard next? Keep track of what we are creating at https://lnkd.in/eMrfi_D3

How To Install Cloudwatch Agent On Amazon Linux By Using CLI Step 1 - Create EC2 and attach CloudWatchAgentAdminPolicy role to EC2 Choose CloudWatchAgentAdminPolicy role which already created by using AWS Managed Rule CloudWatchAgentAdminPolicy For installation CloudWatch agent we need to execute the following command: sudo yum install amazon... https://lnkd.in/es_4VgsP By Thu Kha Kyawe

Amazon CloudWatch Agent adds support for Windows Event Log Filters - Amazon CloudWatch agent has added support for configurable Windows Event log filters. This new feature allows customers to selectively collect and send system and application events to CloudWatch from Windows hosts running on Amazon EC2 or on-premises. The addition of… https://lnkd.in/edPrPFE4

🚀 Amazon EC2 now supports Optimize CPUs for License-Included Instances giving you the ability to fine-tune vCPUs and even disable hyper-threading for Windows Server and SQL Server instances that include licenses. 💰 Why this matters: Most organizations pay for Windows and SQL licenses based on vCPU not actual workload needs. With this update, you can adjust CPU counts to match real performance requirements, cutting licensing costs by up to 50%, while keeping the same memory and storage performance. ⚙️ Example: An EC2 instance with 32 vCPUs can now run efficiently with 16 vCPUs — same memory, same IOPS, half the licensing cost. This reflects AWS’s continued push toward smarter, more flexible infrastructure empowering teams to achieve both cost optimization and operational efficiency. If you’re managing Windows or SQL Server workloads on EC2, this feature is worth exploring. Sometimes, small configuration tweaks lead to big savings. 💡 #AWS #EC2 #CloudComputing #CostOptimization #WindowsServer #SQLServer #CloudEngineering #AWSCommunity

Boring Thursday to You! It is not boring for me in AWS it seems. Couple of interesting news if you are not following AWS: Application Load Balancer now supports JWT validation! Even in service to service communication (glory to the internal ALBs!) - Why is it important? Imagine having a game backend API... and now you have Json Web Token from Oauth 2.0 before you had to use Lambda or API microservice now it is fully managed in ALB. More here: https://lnkd.in/dsF_yv6K Next is the AWS Site-To-Site VPN - remember ECMP and your network speciality certification with 1.25Gbps per tunnel? Not anymore. Now it allows for whooping 5Gbps, there is unfortunately a catch 1.25Gbps costs 0.05$ while 5Gbps (4x) costs 0.6$. Comparing - 5Gbps single tunnel 432$/month 4x1.25Gbps 144$ - I don't like this new math I must say. (https://lnkd.in/dWGVm9sW) Another cool thing is the Network Load Balancer access logs now publishes logs to the CloudWatch (If you build some neat dashboards for access logs now it can be done in CW only) - https://lnkd.in/dEnZkB_D Last thing I found interesting during this pre re:Invent time (remember all product teams keeps big news under the blanket now for 18 more days) is the DCV this is Amazon Desktop streaming (it is free if the desktop is in AWS) supports MAC instances with Apple Silicon - https://lnkd.in/dn82m7CN While gaming prefers Parsec this free alternative is here to help and it constantly improves! That's all folks till the next time (flies away with too many brackets in this post anyway). #aws #news #justcuriousifsomebodyreadsthisornotorsomebodywillbanmeforthislongtags

New AWS Update: Optimize CPUs for License-Included Windows and SQL Server EC2 Instances I am excited about this update. AWS now supports Optimize CPUs for license-included Windows Server and SQL Server EC2 instances, giving customers more flexibility when balancing memory requirements and licensing costs. Many customers have workloads that need large amounts of memory, but do not actually need all the vCPUs that come with a bigger instance size. This often leads to higher Windows and SQL Server licensing costs. This new capability helps address that challenge. Customers can now keep the same memory, storage performance, and network bandwidth from their chosen instance size, and configure a lower number of active vCPUs to better match their workload needs. This also helps optimize licensing costs. For example, an r7i.8xlarge instance with 32 vCPUs, 256 GiB of memory, and 40,000 IOPS can be configured to run with 16 active vCPUs. This reduces SQL Server licensing costs while keeping memory and IOPS unchanged. This is great news for anyone running Windows or SQL Server workloads on EC2, and it is definitely worth exploring. #AWS #EC2 #WindowsServer #SQLServer #CloudOptimization #CostOptimization #AWSWhatsNew https://lnkd.in/gamGvcMn

Have you ever wondered what instance types you should be using in AWS? It's not easy, there are 145 instance families and 1.5 million prices across EC2/RDS. It's become so complex it seems almost impossible to work out what you should use. Not to worry - we have all AWS prices and have run benchmarks on every instance available globally, so you don't have to. Here’s what you need to know. Key Takeaways: 🏆 The Undisputed Winner: The m8a family (AMD) provides incredible performance for the price. The m8a delivers both the highest raw CPU performance (310% compared to the m8i) and the absolute best performance-per-dollar found on AWS. As the m8a is available only in some regions, the m7a is the next best option with 80% of the m8a's performance at a 5% lower cost. 💰 The Value King: t4g (Graviton 2). If you're focused purely on cost-efficiency and your workload runs on ARM, the t4g offers incredible value, significantly outperforming every instance type per dollar. If you need consistent high CPU performance and can run ARM, the m6g (Graviton 3) is a close runner-up. 📉 What to Avoid: You should immediately re-evaluate using older-generation CPUs (m5/m5a) or even newer Intel-based (m6i, m7i) and flex instances. For general-purpose workloads, they offer significantly lower value, with some providing less than 1/4th the performance-per-dollar of the newer AMD series. Our Recommendations: ⚡ For Maximum Performance: Your new default should be the m7a or m8a (AMD) families. 💰 For Maximum Cost-Efficiency: Your default should be Graviton, either the t4g or m6g series. ⚙️ The Best All-Rounder: The m6a family offers the best balance of great performance and low cost, making it the ideal choice for most workloads. 🧠 The Bottom Line: If you are still using m5, t3, or m6i instances, you are likely overspending. To see for yourself, our free public pricing calculator has all EC2/RDS costs, with SPs, RIs, and sysbench results for all AWS instance families. cloudhound.io For AWS partners, CloudHound is the market leader for AWS assessment tooling and is fully funded for MAP and OLA assessments. #AWS #CloudComputing #EC2 #FinOps #CloudCosts #AMD #Graviton #Benchmarking

Have you checked your running EC2 instance types lately? A simple spot-check of your instances to ensure they're on latest, greatest generations can lead to cost savings and/or better performance. If you're on older generations, you're spending more and getting less.