AWS Lambda functions scale up 12 times faster

AWS Lambda functions scale up 12 times faster

Updates related to Cloud TPU, Filestore, AKS, Dynamic Workload Scheduler, Azure NetApp, Gemini and more

NPW Research

Orchestration

AWS Lambda functions scale up 12 times faster

  • AWS Lambda now scales 12 times faster for high-volume requests.
  • Independent scaling of functions within an account.
  • No additional cost or configuration required. UPDATE

Filestore enhances GKE stateful workloads

  • Filestore now supports high-capacity Zonal.
  • Introduces CSI-integrated backups, multi-shares.
  • Offers high durability, performance. UPDATE

AKS now supports 5K node limit

Azure Kubernetes Service (AKS) now allows scaling up to 5000 nodes in Standard tier clusters, enhancing scalability. UPDATE

Provisioning + Runtime

Google Cloud introduces TPU v5p, AI Hypercomputer

  • Cloud TPU v5p for generative AI models.
  • AI Hypercomputer integrates optimized AI stack.
  • Designed for AI workload efficiency. UPDATE

Dynamic Workload Scheduler for AI/ML resources

  • Manages AI/ML resources, optimizing costs.
  • Two modes: Flex Start and Calendar.
  • Supports TPUs, NVIDIA GPUs. UPDATE

Azure NetApp Files supports 1 TiB pools

Azure NetApp Files now allows creating 1 TiB capacity pools, offering more flexibility and cost savings. UPDATE

Google Cloud improves Data Transfer SLA

Google Cloud's Standard Tier Data Transfer now offers a 99.9% availability SLA, enhancing reliability and performance. UPDATE

Security

Microsoft Defender for APIs now available

  • Defender for APIs offers full lifecycle API protection.
  • Integrates with Microsoft CSPM for enhanced security.
  • Supports API attack path analysis. UPDATE

Cloud Armor now for regional load balancers

  • Cloud Armor supports Regional External Load Balancers.
  • Enhances DDoS, web application defense.
  • Aligns with data sovereignty, compliance needs. UPDATE

AI/ML

Introducing Gemini, Google's advanced AI model

  • Gemini: Google's most capable AI model.
  • Offers multimodal capabilities, efficient performance.
  • Supports various sizes: Ultra, Pro, Nano. UPDATE

SageMaker HyperPod reduces training time

  • Reduces foundation model training time by 40%.
  • Supports GPU-based and Trainium-based instances.
  • Enables uninterrupted training with SageMaker libraries. UPDATE

GenAI Databases Retrieval App now open source

  • Open-source app extends LLMs with cloud databases.
  • Reduces hallucinations, improves accuracy.
  • Enhances LLM applications, data access. UPDATE

Observability

Amazon CloudWatch Application Signals in preview

  • CloudWatch Application Signals automates AWS app instrumentation.
  • Provides standardized metrics, pre-built dashboards.
  • Enables easy root cause identification. UPDATE

Azure Monitor releases OpenTelemetry Distro

Azure Monitor's new OpenTelemetry-based Distro for ASP.NET Core Applications is now generally available. UPDATE

NPW Insights this week

  1. Google’s internal code review tool
  2. Shopiy’s developer survey findings
  3. AWS CTO’s 2024 predictions
Access NPW Insights here