Malicious PyPI Packages Stole Cloud Tokens—Over 14,100 Downloads Before Removal

~ Reporter ~ Leave a comment

Cybersecurity researchers have warned of a malicious campaign targeting users of the Python Package Index (PyPI) repository with bogus libraries masquerading as “time” related utilities, but harboring hidden functionality to steal sensitive data such as cloud access tokens.
Software supply chain security firm ReversingLabs said it discovered two sets of packages totaling 20 of them. The packages

from The Hacker News https://ift.tt/KZ5NgDU
via IFTTT

AWS Pi Day 2025: Data foundation for analytics and AI

~ Reporter ~ Leave a comment

Every year on March 14 (3.14), AWS Pi Day highlights AWS innovations that help you manage and work with your data. What started in 2021 as a way to commemorate the fifteenth launch anniversary of Amazon Simple Storage Service (Amazon S3) has now grown into an event that highlights how cloud technologies are transforming data management, analytics, and AI.

This year, AWS Pi Day returns with a focus on accelerating analytics and AI innovation with a unified data foundation on AWS. The data landscape is undergoing a profound transformation as AI emerges in most enterprise strategies, with analytics and AI workloads increasingly converging around a lot of the same data and workflows. You need an easy way to access all your data and use all your preferred analytics and AI tools in a single integrated experience. This AWS Pi Day, we’re introducing a slate of new capabilities that help you build unified and integrated data experiences.

The next generation of Amazon SageMaker: The center of all your data, analytics, and AI
At re:Invent 2024, we introduced the next generation of Amazon SageMaker, the center of all your data, analytics, and AI. SageMaker includes virtually all the components you need for data exploration, preparation and integration, big data processing, fast SQL analytics, machine learning (ML) model development and training, and generative AI application development. With this new generation of Amazon SageMaker, SageMaker Lakehouse provides you with unified access to your data and SageMaker Catalog helps you to meet your governance and security requirements. You can read the launch blog post written by my colleague Antje to learn more details.

Core to the next generation of Amazon SageMaker is SageMaker Unified Studio, a single data and AI development environment where you can use all your data and tools for analytics and AI. SageMaker Unified Studio is now generally available.

SageMaker Unified Studio facilitates collaboration among data scientists, analysts, engineers, and developers as they work on data, analytics, AI workflows, and applications. It provides familiar tools from AWS analytics and artificial intelligence and machine learning (AI/ML) services, including data processing, SQL analytics, ML model development, and generative AI application development, into a single user experience.

SageMaker Unified Studio

SageMaker Unified Studio also brings selected capabilities from Amazon Bedrock into SageMaker. You can now rapidly prototype, customize, and share generative AI applications using foundation models (FMs) and advanced features such as Amazon Bedrock Knowledge Bases, Amazon Bedrock Guardrails, Amazon Bedrock Agents, and Amazon Bedrock Flows to create tailored solutions aligned with your requirements and responsible AI guidelines all within SageMaker.

Last but not least, Amazon Q Developer is now generally available in SageMaker Unified Studio. Amazon Q Developer provides generative AI powered assistance for data and AI development. It helps you with tasks like writing SQL queries, building extract, transform, and load (ETL) jobs, and troubleshooting, and is available in the Free tier and Pro tier for existing subscribers.

You can learn more about SageMaker Unified Studio in this recent blog post written by my colleague Donnie.

During re:Invent 2024, we also launched Amazon SageMaker Lakehouse as part of the next generation of SageMaker. SageMaker Lakehouse unifies all your data across Amazon S3 data lakes, Amazon Redshift data warehouses, and third-party and federated data sources. It helps you build powerful analytics and AI/ML applications on a single copy of your data. SageMaker Lakehouse gives you the flexibility to access and query your data in-place with Apache Iceberg–compatible tools and engines. In addition, zero-ETL integrations automate the process of bringing data into SageMaker Lakehouse from AWS data sources such as Amazon Aurora or Amazon DynamoDB and from applications such as Salesforce, Facebook Ads, Instagram Ads, ServiceNow, SAP, Zendesk, and Zoho CRM. The full list of integrations is available in the SageMaker Lakehouse FAQ.

Building a data foundation with Amazon S3
Building a data foundation is the cornerstone of accelerating analytics and AI workloads, enabling organizations to seamlessly manage, discover, and utilize their data assets at any scale. Amazon S3 is the world’s best place to build a data lake, with virtually unlimited scale, and it provides the essential foundation for this transformation.

I’m always astonished to learn about the scale at which we operate Amazon S3: It currently holds over 400 trillion objects, exabytes of data, and processes a mind-blowing 150 million requests per second. Just a decade ago, not even 100 customers were storing more than a petabyte (PB) of data on S3. Today, thousands of customers have surpassed the 1 PB milestone.

Amazon S3 stores exabytes of tabular data, and it averages over 15 million requests to tabular data per second. To help you reduce the undifferentiated heavy lifting when managing your tabular data in S3 buckets, we announced Amazon S3 Tables at AWS re:Invent 2024. S3 Tables are the first cloud object store with built-in support for Apache Iceberg. S3 tables are specifically optimized for analytics workloads, resulting in up to threefold faster query throughput and up to tenfold higher transactions per second compared to self-managed tables.

Today, we’re announcing the general availability of Amazon S3 Tables integration with Amazon SageMaker Lakehouse  Amazon S3 Tables now integrate with Amazon SageMaker Lakehouse, making it easy for you to access S3 Tables from AWS analytics services such as Amazon Redshift, Amazon Athena, Amazon EMR, AWS Glue, and Apache Iceberg–compatible engines such as Apache Spark or PyIceberg. SageMaker Lakehouse enables centralized management of fine-grained data access permissions for S3 Tables and other sources and consistently applies them across all engines.

For those of you who use a third-party catalog, have a custom catalog implementation, or only need basic read and write access to tabular data in a single table bucket, we’ve added new APIs that are compatible with the Iceberg REST Catalog standard. This enables any Iceberg-compatible application to seamlessly create, update, list, and delete tables in an S3 table bucket. For unified data management across all of your tabular data, data governance, and fine-grained access controls, you can also use S3 Tables with SageMaker Lakehouse.

To help you access S3 Tables, we’ve launched updates in the AWS Management Console. You can now create a table, populate it with data, and query it directly from the S3 console using Amazon Athena, making it easier to get started and analyze data in S3 table buckets.

The following screenshot shows how to access Athena directly from the S3 console.

S3 console : create table with AthenaWhen I select Query tables with Athena or Create table with Athena, it opens the Athena console on the correct data source, catalog, and database.

S3 Tables in Athena

Since re:Invent 2024, we’ve continued to add new capabilities to S3 Tables at a rapid pace. For example, we added schema definition support to the CreateTable API and you can now create up to 10,000 tables in an S3 table bucket. We also launched S3 Tables into eight additional AWS Regions, with the most recent being Asia Pacific (Seoul, Singapore, Sydney) on March 4, with more to come. You can refer to the S3 Tables AWS Regions page of the documentation to get the list of the eleven Regions where S3 Tables are available today.

Amazon S3 Metadata—announced during re:Invent 2024— has been generally available since January 27. It’s the fastest and easiest way to help you discover and understand your S3 data with automated, effortlessly-queried metadata that updates in near real time. S3 Metadata works with S3 object tags. Tags help you logically group data for a variety of reasons, such as to apply IAM policies to provide fine-grained access, specify tag-based filters to manage object lifecycle rules, and selectively replicate data to another Region. In Regions where S3 Metadata is available, you can capture and query custom metadata that is stored as object tags. To reduce the cost associated with object tags when using S3 Metadata, Amazon S3 reduced pricing for S3 object tagging by 35 percent in all Regions, making it cheaper to use custom metadata.

AWS Pi Day 2025
Over the years, AWS Pi Day has showcased major milestones in cloud storage and data analytics. This year, the AWS Pi Day virtual event will feature a range of topics designed for developers and technical decision-makers, data engineers, AI/ML practitioners, and IT leaders. Key highlights include deep dives, live demos, and expert sessions on all the services and capabilities I discussed in this post.

By attending this event, you’ll learn how you can accelerate your analytics and AI innovation. You’ll learn how you can use S3 Tables with native Apache Iceberg support and S3 Metadata to build scalable data lakes that serve both traditional analytics and emerging AI/ML workloads. You’ll also discover the next generation of Amazon SageMaker, the center for all your data, analytics, and AI, to help your teams collaborate and build faster from a unified studio, using familiar AWS tools with access to all your data whether it’s stored in data lakes, data warehouses, or third-party or federated data sources.

For those looking to stay ahead of the latest cloud trends, AWS Pi Day 2025 is an event you can’t miss. Whether you’re building data lakehouses, training AI models, building generative AI applications, or optimizing analytics workloads, the insights shared will help you maximize the value of your data.

Tune in today and explore the latest in cloud data innovation. Don’t miss the opportunity to engage with AWS experts, partners, and customers shaping the future of data, analytics, and AI.

If you missed the virtual event on March 14, you can visit the event page at any time—we will keep all the content available on-demand there!

— seb

How is the News Blog doing? Take this 1 minute survey!

(This survey is hosted by an external company. AWS handles your information as described in the AWS Privacy Notice. AWS will own the data gathered via this survey and will not share the information collected with survey respondents.)

from AWS News Blog https://ift.tt/UlkP8iM
via IFTTT

Guardz Reveals Details of an Ongoing Phishing Campaign Exploiting Microsoft 365 Infrastructure

~ Reporter ~ Leave a comment

The cybersecurity company empowering MSPs to secure small businesses identified a highly sophisticated Microsoft 365 tenant brand manipulation and disrupted its use against their customers.

Guardz, the cybersecurity company empowering MSPs and IT professionals to deliver comprehensive, AI-native cyber protection for small businesses, today disclosed the findings of its research into a highly sophisticated, ongoing phishing campaign exploiting Microsoft 365’s trusted infrastructure to manipulate victims into calling a malicious threat actor call center and potentially facilitate credential harvesting and account takeover (ATO) attempts.

As email security defenses like secure email gateways (SEGs) and advanced threat protection mechanisms become more complex, cyber threat actors are continuously refining their evasion techniques to bypass even the most robust detection mechanisms. Evidencing this trend, Guardz identified, analyzed, and successfully disrupted a highly deceptive phishing campaign in use against its customers, highlighting how cyber attackers continue to evolve their techniques by manipulating legitimate infrastructure in novel ways.

The Guardz Research Unit (GRU) has determined the details of the attack method, which exploits legitimate Microsoft services to create a trusted delivery mechanism for phishing content, making it difficult for both technical controls and human recipients to detect it. By manipulating Microsoft 365 tenant properties and leveraging organizational profile spoofing to embed phishing payloads directly within legitimate emails, attackers are able to trick users into providing information under the cloak of legitimacy.

The attack flow involves numerous phases:

●Infrastructure Acquisition: Adversaries establish control over multiple Microsoft 365 organization tenants, either by registering new tenants or compromising existing ones. Each tenant plays a strategic role in the attack chain, allowing the threat actor to evade detection and manipulate trust mechanisms within the Microsoft 365 infrastructure. This can allow various attack functionalities, including exploiting legitimate payment and billing activity emails sent by Microsoft with phishing content.

●Technical Configuration: Once the control over Microsoft 365 tenants is established, the attacker can create administrative accounts using the default “*.onmicrosoft.com” domain. The key tactics include admin account creation, mail forwarding abuse, and anti-phishing evasion.

●Deception Preparation: To enhance the credibility of their phishing campaign, attackers configure the second tenant’s organization name with a misleading full-text message that mimics a legitimate Microsoft transaction notification. This tactic exploits Microsoft 365’s built-in tenant display name feature, which is reflected in various service-generated emails and interfaces, to inject a phishing lure directly into the email.

● Attack Execution: To maximize legitimacy and evade detection, the attacker initiates a purchase or trial subscription event within the first tenant. This action generates an authentic Microsoft-signed billing email, leveraging Microsoft’s infrastructure to deliver phishing content that appears completely legitimate. The attacker manipulates the organization display name in a second tenant, ensuring that the fraudulent message is embedded within a trusted communication channel. Because the emails leverage native M365 infrastructure and the sending domain is legitimately Microsoft.com, the phishing lures cannot be detected by SPF, DKIM, and DMARC.

●Technical Legitimization: By leveraging Microsoft’s legitimate email infrastructure, the attacker ensures that the phishing email passes through Microsoft’s servers without raising security alerts. Because the email originates from a trusted source, it is far more likely to reach the victim’s inbox without being flagged by security tools.

●Victim Engagement: Microsoft’s billing emails contain the organization name and fake support contact numbers, urging immediate victim interaction with a call center. This direct communication significantly enhances phishing effectiveness beyond traditional email-based methods.

“Our team at Guardz works tirelessly to secure small businesses, who are the backbone of the US economy and who threat actors are increasingly setting their sights on – and we’re proud to have identified and protected against this highly deceptive attack,” said Dor Eisner, CEO and Co-Founder of Guardz. “By exploiting the inherent trust in Microsoft’s cloud services, this phishing campaign is significantly more challenging for security teams to detect and mitigate, evading domain reputation analysis, DMARC enforcement, and anti-spoofing mechanisms. It’s an urgent reminder that as cyber defenders, we must focus not only on traditional indicators of compromise but also on how legitimate systems can be manipulated for malicious purposes.”

The Guardz unified security platform gives a unique edge in combating this type of threat. The company’s unified detection and response effectively mitigated the attack, while its security team informed affected customers and implemented enhanced detection mechanisms to prevent similar threats in the future.

To protect against this attack vector, Guardz recommends that businesses implement enhanced detection and response tools, starting with email analysis that includes advanced content inspection, user awareness training, phone verification validating official support numbers, and verification of unknown domains and newly created tenants.

To learn more about the phishing campaign and how Guardz protects against it, read the full blog post here.

__

 About Guardz

Guardz provides MSPs and IT professionals with an AI-powered cybersecurity platform designed to secure and insure SMBs against cyberattacks. The Guardz platform offers automatic detection and response, protecting users, emails, devices, cloud directories, and data. By simplifying cybersecurity management, Guardz enables businesses to focus on growth without being bogged down by security complexities. The company’s scalable and cost-effective pricing model ensures comprehensive protection for all digital assets, facilitating rapid deployment and business expansion.

The post Guardz Reveals Details of an Ongoing Phishing Campaign Exploiting Microsoft 365 Infrastructure appeared first on Cybersecurity Insiders.

from Cybersecurity Insiders https://ift.tt/mLNAzhP
via IFTTT

The DevOps Threats Unwrapped: over 502 incidents and 955 hours of disruptions in GitHub, GitLab, Atlassian, and Azure DevOps

~ Reporter ~ Leave a comment

502 incidents, including 48 at the highest risk level, resulting in a total of 955 hours of major and critical disruptions – that’s 120 business days… These are the conclusions of The DevOps Threats Unwrapped report prepared by the GitProtect research team, which analyzed all GitHub, GitLab, Atlassian, and Azure DevOps incidents over the past year.

2024 was a wake-up call for the world of data protection with the Crowdstrike-Microsoft incident serving as a stark reminder of the growing severity of security breaches. This event underscored the vulnerability of even the most robust organizations, leaving a trail of devastation: $5.4 billion in damages and 8.5 million affected Windows devices worldwide.

However, these headline-grabbing figures are just the tip of the iceberg. The year marked a surge in attacks targeting SaaS applications and DevOps tools, a trend that will accelerate in 2025.

GitProtect.io, the world’s most trusted DevOps backup and disaster recovery vendor, launched its new edition of The DevOps Threats Unwrapped study. The research reveals the most severe flaws, prolonged outages, devastating human errors, data breaches, and other incidents that shaped the DevOps cybersecurity landscape last year. The study focuses on GitHub, GitLab, Bitbucket, Jira, and Azure DevOps data protection. 

According to research, in 2024 DevOps had to handle 502 incidents impacting those tools, including 48 with the highest level of risk. Also, according to statistics, they struggled with interruptions in availability – 955 hours of major and critical disruptions in total. It gives… nearly 120 working days a year(!). Add to this the incidents that caused temporary interruptions in availability or degraded performance and the scale becomes enormous.

To give this a perspective, 955 hours is enough to cross the Atlantic Ocean from Europe by small yacht, with a short break in the Caribbean, reach the East Coast, and… go back to Portugal. 

Breaking these numbers down by platform, GitHub users suffered 124 incidents including 26 with major impact that caused 134 hours of disruptions. Bitbucket noted 38 incidents with 4 critical, responsible for 4 hours of disruption. Jira – 132 incidents, 10 critical ones resulting in 17 hours of outage. Azure DevOps – 111 incidents, 1 with unhealthy status that lasted almost 2 hours. Finally, GitLab with 97 issues, 7 service disruption incidents, and a record amount of 798 hours of disruption.

Beyond Vendor Outages: A Growing Threat Landscape

Service disruptions are only part of the challenge. Among the most pervasive threats to DevOps continuity and data integrity are hardcoded secrets, unsecured databases, repo jacking, intruders in the software supply chain, the growing scale of AI-generated threats, and unchanged – various types of human errors.

According to the GitProtect.io study, last year alone, brands such as, among others, Mercedes, New York Times, Schneider Electric, Cisco, the Chinese Ministry of Public Security, and Cloudflare have announced incidents involving hacking or breaches of their GitHub, GitLab, Atlassian stack or AzureDevOps data. Descriptions of these incidents are available on The DevOps Threats Unwrapped report page.

The top 3 impacted industries were Technology and Software, Fintech/Banking, Media, and Entertainment. Healthcare, government entities, telecommunications, and manufacturing also made the shortlist.

Shared Responsibility Models and Growing Compliance Requirements

In defense of SaaS vendors, the Shared Responsibility model, which is used by virtually all SaaS and cloud providers, operates. It assumes shared responsibility in many areas of cybersecurity. GitHub, GitLab, Atlassian, and Microsoft all inform their customers about the need to fulfill security obligations, including mandatory DevOps data backup on the account level.

What they must focus on this year, is greater and more intensive education on user responsibilities and the need to take care of data security and backup.

Fortunately, this growing scale of threats is already making SaaS customers aware of the need to secure their data. According to Gartner, by 2028 75% of enterprises will prioritize backup of SaaS apps as a critical requirement, compared to 15% in 2024. 

This trend also does not escape the attention of legislation and government institutions. 2024 also saw a rising interest in compliance and regulatory issues. Just to mention Digital Operational Resilience Act (DORA) that came into force on the 17th of February, NIS 2, SOC 2, HIPAA, and more security acts dedicated to the most critical and at the same time, vulnerable sectors. These frameworks mandate robust data protection measures, including backups of git repositories, Jira projects, and other DevOps tools.

Looking Ahead: Predictions and Best Practices for 2025

As cyber threats continue to evolve, organizations must remain vigilant. The 2024 DevSecOps Threats Unwrapped report by GitProtect.io provides actionable insights and forecasts for 2025, empowering enterprises to bolster their data security strategies.

For a deeper dive into these findings and expert recommendations, visit the official report page.

 

__

 

Author – about us

GitProtect.io by Xopero Software is a world-leading automated and manageable backup and Disaster Recovery solution for all Jira, Bitbucket, GitHub, GitLab, Azure DevOps and more stack data. It ensures DevOps with data accessibility and seamless workflow, even during service downtime. Trusted by Security Teams, it helps to meet the Cloud Shared Responsibility Model, comply with security standards (i.e. ISO 27001 or SOC 2) and empower them with audit-ready governance, advanced reporting, and best-in-class security controls.

More information: https://gitprotect.io/

 

 

 

The post The DevOps Threats Unwrapped: over 502 incidents and 955 hours of disruptions in GitHub, GitLab, Atlassian, and Azure DevOps appeared first on Cybersecurity Insiders.

from Cybersecurity Insiders https://ift.tt/RaprzjK
via IFTTT

Accelerating Mergers and Acquisitions with Zero Trust Network Access (ZTNA)

~ Reporter ~ Leave a comment

Introduction

Mergers and acquisitions (M&A) are critical growth strategies for businesses, but they come with significant IT and security challenges. A smooth transition requires the rapid integration of networks, secure access to applications, and the protection of sensitive data. 

Traditional security models, which assume trust within a corporate network, often struggle to meet these demands efficiently. Zero Trust Network Access (ZTNA) offers a modern, scalable, and secure approach to accelerate M&A by enabling seamless access management, reducing cyber risks, and ensuring business continuity.

Challenges of IT Integration in M&A

During M&A, organizations face several IT-related obstacles, including:

  • Complex Network Mergers: Companies often have different network architectures, requiring extensive reconfiguration.
  • Security Vulnerabilities: Integrating disparate IT environments can create security gaps, increasing the attack surface for cyber threats.
  • Access Management Difficulties: Ensuring secure and role-based access to critical resources across organizations is challenging.
  • Compliance and Regulatory Requirements: Companies must align their security frameworks with regulatory standards such as GDPR, NIS2, and industry-specific guidelines.
  • Operational Disruptions: IT integration delays can impact productivity, delaying the overall success of the M&A.

How ZTNA Accelerates M&A

ZTNA provides a robust security framework that simplifies IT integration, enhances security, and ensures seamless access to critical systems. Below are key ways ZTNA accelerates the M&A process:

Seamless and Secure User Onboarding

ZTNA enables fast and secure onboarding of employees from both merging organizations by:

  • Identity-Centric Access Control: Ensuring that only authorized users access specific applications and data.
  • Multi-Factor Authentication (MFA): Enhancing security by requiring additional identity verification.
  • Cloud-Native Deployment: Eliminating the need for complex on-premises infrastructure, allowing rapid provisioning.

Enhanced Security with Zero Trust Principles

Unlike traditional perimeter-based security, ZTNA follows a ‘never trust, always verify’ approach by:

  • Continuous User and Device Verification: Ensuring that each access request is assessed in real time.
  • Microsegmentation: Restricting lateral movement within the network to minimize the risk of breaches.
  • Context-Aware Access Policies: Dynamically adjusting access based on user identity, location, device health, and real-time risk analysis.

Regulatory Compliance and Risk Reduction

ZTNA ensures compliance with various regulatory frameworks, such as:

  • Data Encryption: Protecting sensitive information in transit and at rest.
  • Granular Access Logs: Providing detailed records for auditing and compliance reporting.
  • Least Privilege Access Enforcement: Minimizing exposure by granting access based on necessity and context.

Ensuring Business Continuity and Productivity

M&A success depends on seamless business operations. ZTNA supports:

  • Hybrid and Remote Work Enablement: Allowing employees to securely access corporate applications from any location.
  • Scalability and Flexibility: Adapting to the growing needs of the merged entity without major infrastructure changes.
  • Reduced Downtime: Streamlining integration processes and minimizing disruptions.

Conclusion

ZTNA is a game-changer for accelerating M&A by providing a secure, scalable, and efficient approach to IT integration. By adopting a Zero Trust strategy, organizations can ensure seamless user onboarding, protect sensitive assets, maintain regulatory compliance, and minimize operational disruptions. 

In today’s fast-evolving digital landscape, leveraging ZTNA is not just an advantage—it’s essential for a successful and secure M&A process.

 

The post Accelerating Mergers and Acquisitions with Zero Trust Network Access (ZTNA) appeared first on Cybersecurity Insiders.

from Cybersecurity Insiders https://ift.tt/eJbxGQ7
via IFTTT

Collaborate and build faster with Amazon SageMaker Unified Studio, now generally available

~ Reporter ~ Leave a comment

Today, we’re announcing the general availability of Amazon SageMaker Unified Studio, a single data and AI development environment where you can find and access all of the data in your organization and act on it using the best tool for the job across virtually any use case. Introduced as preview during AWS re:Invent 2024, my colleague, Antje, summarized it as:

SageMaker Unified Studio (preview) is a single data and AI development environment. It brings together functionality and tools from the range of standalone “studios,” query editors, and visual tools that we have today in Amazon Athena, Amazon EMR, AWS Glue, Amazon Redshift, Amazon Managed Workflows for Apache Airflow (Amazon MWAA), and the existing SageMaker Studio.

Here’s a video to see Amazon SageMaker Unified Studio in action:

SageMaker Unified Studio breaks down silos in data and tools, giving data engineers, data scientists, data analysts, ML developers and other data practitioners a single development experience. This saves development time and simplifies access control management so data practitioners can focus on what really matters to them—building data products and AI applications.

This post focuses on several important announcements that we’re excited to share:

To get started, go to the Amazon SageMaker console and create a SageMaker Unified Studio domain. To learn more, visit Create an Amazon SageMaker Unified Studio domain in the AWS documentation.

New capabilities for Amazon Bedrock in SageMaker Unified Studio
The capabilities of Amazon Bedrock within Amazon SageMaker Unified Studio offer a governed collaborative environment for developers to rapidly create and customize generative AI applications. This intuitive interface caters to developers of all skill levels, providing seamless access to the high-performance FMs offered in Amazon Bedrock and advanced customization tools for collaborative development of tailored generative AI applications.

Since the preview launch, several new FMs have become available in Amazon Bedrock and are fully integrated with SageMaker Unified Studio, including Anthropic’s Claude 3.7 Sonnet and DeepSeek-R1. These models can be used for building generative AI apps and chatting in the playground in SageMaker Unified Studio.

Here’s how you can choose Anthropic’s Claude 3.7 Sonnet on the model selection in your project.

You can also source data or documents from S3 folders within your project and select specific FMs when creating knowledge bases. 

During preview, we introduced Amazon Bedrock Guardrails to help you implement safeguards for your Amazon Bedrock application based on your use cases and responsible AI policies. Now, Amazon Bedrock Guardrails is extended to Amazon Bedrock Flows with this general availability release.

Additionally, we have streamlined generative AI setup for associated accounts with a new user management interface in SageMaker Unified Studio, making it straightforward for domain administrators to grant associated account admins access to model governance projects. This enhancement eliminates the need for command line operations, streamlining the process of configuring generative AI capabilities across multiple AWS accounts.

These new features eliminate barriers between data, tools, and builders in the generative AI development process. You and your team will gain a unified development experience by incorporating the powerful generative AI capabilities of Amazon Bedrock — all within the same workspace.

Amazon Q Developer is now generally available in SageMaker Unified Studio
Amazon Q Developer is now generally available in Amazon SageMaker Unified Studio, providing data professionals with generative AI–powered assistance across the entire data and AI development lifecycle.

Amazon Q Developer integrates with the full suite of AWS analytics and AI/ML tools and services within SageMaker Unified Studio, including data processing, SQL analytics, machine learning model development, and generative AI application development, to accelerate collaboration and help teams build data and AI products faster. To get started, you can select Amazon Q Developer icon.

For new users of SageMaker Unified Studio, Amazon Q Developer serves as an invaluable onboarding assistant. It can explain core concepts such as domains and projects, provide guidance on setting up environments, and answer your questions.

Amazon Q Developer helps you discover and understand data using powerful natural language interactions with SageMaker Catalog. What makes this implementation particularly powerful is how Amazon Q Developer combines broad knowledge of AWS analytics and AI/ML services with the user’s context to provide personalized guidance.

You can chat about your data assets through a conversational interface, asking questions such as “Show all payment related datasets” without needing to navigate complex metadata structures.

Amazon Q Developer offers SQL query generation through its integration with the built-in query editor available in SageMaker Unified Studio. Data professionals of varying skill levels can now express their analytical needs in natural language, receiving properly formatted SQL queries in return.

For example, you can ask, “Analyze payment method preferences by age group and region” and Amazon Q Developer will generate the appropriate SQL with proper joins across multiple tables.

Additionally, Amazon Q Developer is also available to assist with troubleshooting and generating real-time code suggestions in SageMaker Unified Studio Jupyter notebooks, as well as building ETL jobs.

Now available

  • Availability — Amazon SageMaker Unified Studio is now available in the following AWS Regions: US East (N. Virginia, Ohio), US West (Oregon), Asia Pacific (Seoul, Singapore, Sydney, Tokyo), Canada (Central), Europe (Frankfurt, Ireland, London), South America (São Paulo). Learn more about the availability of these capabilities on supported Region documentation page.
  • Amazon Q Developer subscription — The free tier of Amazon Q Developer is available by default in SageMaker Unified Studio, requiring no additional setup or configuration. If you already have Amazon Q Developer Pro Tier subscriptions, you can use those enhanced capabilities within the SageMaker Unified Studio environment. For more information, visit the documentation page.
  • Amazon Bedrock capabilities — To learn more about the capabilities of Amazon Bedrock in Amazon SageMaker Unified Studio, refer to this documentation page. 

Start building with Amazon SageMaker Unified Studio today. For more information, visit the Amazon SageMaker Unified Studio page.

Happy building!

— Donnie Prakoso

— How is the News Blog doing? Take this 1 minute survey! (This survey is hosted by an external company. AWS handles your information as described in the AWS Privacy Notice. AWS will own the data gathered via this survey and will not share the information collected with survey respondents.)

from AWS News Blog https://ift.tt/AHvLWwJ
via IFTTT

Amazon S3 Tables integration with Amazon SageMaker Lakehouse is now generally available

~ Reporter ~ Leave a comment

At re:Invent 2024, we launched Amazon S3 Tables, the first cloud object store with built-in Apache Iceberg support to streamline storing tabular data at scale, and Amazon SageMaker Lakehouse to simplify analytics and AI with a unified, open, and secure data lakehouse. We also previewed S3 Tables integration with Amazon Web Services (AWS) analytics services for you to stream, query, and visualize S3 Tables data using Amazon Athena, Amazon Data Firehose, Amazon EMR, AWS Glue, Amazon Redshift, and Amazon QuickSight.

Our customers wanted to simplify the management and optimization of their Apache Iceberg storage, which led to the development of S3 Tables. They were simultaneously working to break down data silos that impede analytics collaboration and insight generation using the SageMaker Lakehouse. When paired with S3 Tables and SageMaker Lakehouse in addition to built-in integration with AWS analytics services, they can gain a comprehensive platform unifying access to multiple data sources enabling both analytics and machine learning (ML) workflows.

Today, we’re announcing the general availability of Amazon S3 Tables integration with Amazon SageMaker Lakehouse to provide unified S3 Tables data access across various analytics engines and tools. You can access SageMaker Lakehouse from Amazon SageMaker Unified Studio, a single data and AI development environment that brings together functionality and tools from AWS analytics and AI/ML services. All S3 tables data integrated with SageMaker Lakehouse can be queried from SageMaker Unified Studio and engines such as Amazon Athena, Amazon EMR, Amazon Redshift, and Apache Iceberg-compatible engines like Apache Spark or PyIceberg.

With this integration, you can simplify building secure analytic workflows where you can read and write to S3 Tables and join with data in Amazon Redshift data warehouses and third-party and federated data sources, such as Amazon DynamoDB or PostgreSQL.

You can also centrally set up and manage fine-grained access permissions on the data in S3 Tables along with other data in the SageMaker Lakehouse and consistently apply them across all analytics and query engines.

S3 Tables integration with SageMaker Lakehouse in action
To get started, go to the Amazon S3 console and choose Table buckets from the navigation pane and select Enable integration to access table buckets from AWS analytics services.

Now you can create your table bucket to integrate with SageMaker Lakehouse. To learn more, visit Getting started with S3 Tables in the AWS documentation.

1. Create a table with Amazon Athena in the Amazon S3 console
You can create a table, populate it with data, and query it directly from the Amazon S3 console using Amazon Athena with just a few steps. Select a table bucket and select Create table with Athena, or you can select an existing table and select Query table with Athena.

2. Create tables with Athena

When you want to create a table with Athena, you should first specify a namespace for your table. The namespace in an S3 table bucket is equivalent to a database in AWS Glue, and you use the table namespace as the database in your Athena queries.

Choose a namespace and select Create table with Athena. It goes to the Query editor in the Athena console. You can create a table in your S3 table bucket or query data in the table.

2. Query with Athena

2. Query with SageMaker Lakehouse in the SageMaker Unified Studio
Now you can access unified data across S3 data lakes, Redshift data warehouses, third-party and federated data sources in SageMaker Lakehouse directly from SageMaker Unified Studio.

To get started, go to the SageMaker console and create a SageMaker Unified Studio domain and project using a sample project profile: Data Analytics and AI-ML model development. To learn more, visit Create an Amazon SageMaker Unified Studio domain in the AWS documentation.

After the project is created, navigate to the project overview and scroll down to project details to note down the project role Amazon Resource Name (ARN).

3. Project details in SageMaker Unified Studio

Go to the AWS Lake Formation console and grant permissions for AWS Identity and Access Management (IAM) users and roles. In the in the Principals section, select the <project role ARN> noted in the previous paragraph. Choose Named Data Catalog resources in the LF-Tags or catalog resources section and select the table bucket name you created for Catalogs. To learn more, visit Overview of Lake Formation permissions in the AWS documentation.

4. Grant permissions in Lake Formation console

When you return to SageMaker Unified Studio, you can see your table bucket project under Lakehouse in the Data menu in the left navigation pane of project page. When you choose Actions, you can select how to query your table bucket data in Amazon Athena, Amazon Redshift, or JupyterLab Notebook.

5. S3 Tables in Unified Studio

When you choose Query with Athena, it automatically goes to Query Editor to run data query language (DQL) and data manipulation language (DML) queries on S3 tables using Athena.

Here is a sample query using Athena:

select * from "s3tablecatalog/s3tables-integblog-bucket”.”proddb"."customer" limit 10;

6. Athena query in Unified Studio

To query with Amazon Redshift, you should set up Amazon Redshift Serverless compute resources for data query analysis. And then you choose Query with Redshift and run SQL in the Query Editor. If you want to use JupyterLab Notebook, you should create a new JupyterLab space in Amazon EMR Serverless.

3. Join data from other sources with S3 Tables data
With S3 Tables data now available in SageMaker Lakehouse, you can join it with data from data warehouses, online transaction processing (OLTP) sources like relational or non-relational database, Iceberg tables, and other third party sources to gain more comprehensive and deeper insights.

For example, you can add connections to data sources such as Amazon DocumentDB, Amazon DynamoDB, Amazon Redshift, PostgreSQL, MySQL, Google BigQuery, or Snowflake and combine data using SQL without extract, transform, and load (ETL) scripts.

Now you can run the SQL query in the Query editor to join the data in the S3 Tables with the data in the DynamoDB.

Here is a sample query to join between Athena and DynamoDB:

select * from "s3tablescatalog/s3tables-integblog-bucket"."blogdb"."customer", 
              "dynamodb1"."default"."customer_ddb" where cust_id=pid limit 10;

To learn more about this integration, visit Amazon S3 Tables integration with Amazon SageMaker Lakehouse in the AWS documentation.

Now available
S3 Tables integration with SageMaker Lakehouse is now generally available in all AWS Regions where S3 Tables are available. To learn more, visit the S3 Tables product page and the SageMaker Lakehouse page.

Give S3 Tables a try in the SageMaker Unified Studio today and send feedback to AWS re:Post for Amazon S3 and AWS re:Post for Amazon SageMaker or through your usual AWS Support contacts.

In the annual celebration of the launch of Amazon S3, we will introduce more awesome launches for Amazon S3 and Amazon SageMaker. To learn more, join the AWS Pi Day event on March 14.

— Channy

—

How is the News Blog doing? Take this 1 minute survey!

(This survey is hosted by an external company. AWS handles your information as described in the AWS Privacy Notice. AWS will own the data gathered via this survey and will not share the information collected with survey respondents.)

from AWS News Blog https://ift.tt/8gk2xJl
via IFTTT