Azure API Management (APIM) is a cloud-based service for creating, publishing, and managing APIs. It offers policy management, access control, and transformation of requests and responses. Self-hosted gateways extend this capability outside of Azure. APIM integrates with API Center for discovery and governance. This solution addresses challenges of managing distributed APIs efficiently.
Tag: IT cloud computing
Microsoft Entra IDaaS in security operations
This architecture demonstrates integrating Microsoft Entra identity and access capabilities into a zero-trust security strategy focused on cloud services. Gartner predicts rapid cloud market growth, highlighting the need for a shift towards prioritizing user identity security. Microsoft Cybersecurity Reference Architecture (MCRA) outlines how these capabilities integrate with existing architectures and provides potential use cases and components.
N-tier application with Apache Cassandra
This reference architecture illustrates deploying virtual machines and a virtual network for an N-tier application using Apache Cassandra on Linux for the data tier. It includes recommendations for virtual machines, virtual network, application gateway, load balancers, network security groups, Cassandra, jumpbox, scalability, subscription limits, performance efficiency, availability, cost, security, and operational excellence. For more information, visit the link provided.
Hub-spoke network topology with Azure Virtual WAN
This hub-spoke architecture in Azure offers a secure hybrid network, using a virtual hub to connect on-premises networks. It employs Azure Virtual WAN for managed services, reducing operational overhead and costs. The architecture ensures improved security and separation of IT concerns. It supports various use cases and offers scalability, reliability, performance, and cost optimization.
SAP deployment in Azure using an Oracle database
This reference architecture outlines best practices for hosting a high-availability SAP NetWeaver with Oracle Database on Azure. It details the components, networking, virtual machines, storage, high availability setup, disaster recovery, backup methods, and considerations for both Windows and Linux deployments. It emphasizes customization based on business requirements and SAP product licensing.
Deploy IBM Maximo Application Suite (MAS) on Azure
IBM Maximo Application Suite (MAS) 8.x runs on OpenShift on Azure. The architecture features a container hosting platform with privatized deployment of worker and control nodes integrated with Azure Premium Files and standard files for storage. MAS includes Manage, Monitor, Health, Visual Inspection, Predict, Assist, Safety, and Civil applications. Deployment on Azure is suggested, and installation is detailed.
Partitioning an LLM between cloud and edge
Large language models (LLMs) have conventionally relied on centralized systems due to high computational demands. However, a partitioned architecture can effectively balance tasks between edge and cloud servers, reducing latency and conserving energy. Despite its complexity, this hybrid approach offers enhanced performance and security for AI deployment, urging businesses to consider its potential.
Let’s Architect! Learn About Machine Learning on AWS
Businesses can enhance operational efficiency and decision-making by adopting a data-driven approach with machine learning (ML). Leveraging ML on AWS can jump-start a data-driven journey, enable MLOps engineering, and implement generative AI infrastructure. Pinterest and Booking.com showcase successful implementations. Amazon SageMaker Immersion Day offers a comprehensive ML training workshop, empowering users to harness AWS ML services.
Creating an organizational multi-Region failover strategy
AWS Regions provide fault isolation boundaries to contain the impact of service impairments, enabling the development of multi-Region applications. Organizations can choose from four high-level strategies for failover: Component-level, Individual application, Dependency graph, and Entire portfolio failover. Each strategy has tradeoffs and requires intentional decision-making for multi-Region failover solutions.
Azure OpenAI chat baseline architecture in an Azure landing zone
This article is part of a series on Azure OpenAI Service, describing the architecture of a generative AI workload deployed in an Azure application landing zone subscription. It details the workload and platform team’s responsibilities, resource ownership, and management, including considerations for networking, data encryption, cost optimization, and operational excellence. For the full article, visit: https://learn.microsoft.com/en-us/azure/architecture/ai-ml/architecture/azure-openai-baseline-landing-zone
Leading Container Security Services for Cloud-Native Environments
Leading Container Security Services In today’s rapidly evolving digital landscape, container security has become a critical component of any organization’s cybersecurity strategy. As enterprises increasingly
Developing a RAG solution – Information retrieval phase
The content provides guidance on generating embeddings, configuring the search index, and experimenting with different searches for information retrieval, including vector, full text, hybrid searches, and filtering. It covers topics such as vector search algorithm, search types, manual multiple queries, and search evaluation methods. It emphasizes experimentation to find the right approach.
Developing a RAG solution – LLM end to end evaluation phase
This content discusses the evaluation of a Retrieval-Augmented Generation (RAG) solution, focusing on metrics like groundedness, completeness, utilization, and relevancy. It also mentions various similarity and evaluation metrics, advising the documentation of hyperparameters and results for future evaluations. The RAG Experiment Accelerator tool is introduced as a means to optimize and enhance the development of RAG solutions.
Developing a RAG solution – Preparation phase
The first phase of Retrieval-Augmented Generation (RAG) development involves preparing the business domain and gathering relevant documents and sample questions. It’s crucial to ensure document pertinence, representation, and quality. Test queries and their outputs are also gathered, and synthetic questions can be generated from representative documents. Unaddressed queries must also be considered. The article provides guidance on document analysis and offers next steps. Source: https://learn.microsoft.com/en-us/azure/architecture/ai-ml/guide/rag/rag-preparation-phase