About this episode
Welcome to episode 272 of The Cloud Pod! This week, Matthew and Justin are bringing you all the latest in cloud and AI news, including new updates to the ongoing Crowdstrike drama, JSON schemas, AWS vaults, and IPv6 addresses – even some hacking opportunities! All this and more, this week in the cloud. Titles we almost went with this week: ️The cloud pod is now logically air-gapped The Cloud Pod has continuous snark The Cloud Pod points the finger at delta AI now with JSON SCHEMAS!!! A big thanks to this week’s sponsor: We’re sponsorless! Want to get your brand, company, or service in front of a very enthusiastic group of cloud news seekers? You’ve come to the right place! Send us an email or hit us up on our slack channel for more info. Follow Up 00:35 Crowdstrike RCA The final RCA is out from Crowdstrike, and as we talked during the preliminary report , this was an issue with a channel file that had 21 input parameters. No update previously had more than 20, and it was not caught in earlier testing. Crowdstrike has several findings, and mitigating actions that they are taking. They go into detail on each of them, and you can read through all of them at the linked document . 02:31 Justin – “…the one thing I would say is this would be a perfect RCA if it included a timeline, but it lacks, it lacks a timeline view.” 12:06 Justin – “…their mitigations don’t have any dates on them of when they’re going to be done or implemented, which, in addition to a timeline, it would be nice to see in this process.” 15:46 Microsoft joins CrowdStrike in pushing IT outage recovery responsibility back to Delta Microsoft has joined Crowdstrike in throwing Delta under the bus. Delta Airlines has been blaming Crowdstrike and MS for their recent IT woes, which the company claims cost them over $ 500 million . Microsoft says “ Our preliminary review suggests that Delta, unlike its competitors, has not modernized its IT infrastructure, either for the benefit of its customers or for its pilots and flight attendants ” Mark Cheffo from law firm Dechert representing MS. Gonna get ugly before this all gets settled. *Insert Michael Jackson eating popcorn gif here* 16:43 Justin – “The struggle with, you know, offering to send someone on site to help you is, you know, you, you can’t vet them that quickly. And so you also have an obligation to your shareholders. You have obligations to your security controls and your SOC and ISO and all the things that you’re doing, you know, to, to allow some strangers into your network and then give them access required to fix this issue, which in some cases required you to provide local encryption keys, and local administrator passwords, like you’re, you’re basically saying, you know, here’s the keys. Cause we’re in a, you know, everything’s in crisis and we’re going to throw security out the window to allow these people to come in and touch my environment to get us back up and running. I could see, I can see the argument both ways.” AI Is Going Great – Or How ML Makes All It’s Money 20:16 Anthropic Offers $15,000 to Break New AI Safety System With Defcon occurring this week, Anthropic is poking the hackers, offering up to $15,000 for “jailbreaks” that bypass the Anthropic AI safeguard and elicit prohibited content from the Claude chatbots. By inviting outside researchers to test the models, Anthropic is hoping to identify problems the company couldn’t find on its own. Anthropic is hoping to attract hacker groups who post jailbreaks on Twitter to recruit for the program. 21:14 Announcing the Generative AI World Cup: A Global Hackathon by Databricks Databricks is hosting a worldwide generative AI world cup hackathon , inviting participants to develop innovative gen AI applications that solve real-world problems. Participants will compete for a pool of over $50,000 in cash prizes, trophies and passes for the Data and AI summit 2025. Participants will also get material and training to help skill up on Generative AI as part of the process. To participate you must meet the eligibility requirements: Participants must hold a data or AI role in their org Register with a corporate email address Teams of 2 to 4 members Databricks staff, partners, consultants, and students are not eligible. Participants must be 20 years old or above, and reside in one of the eligible countries . Deadline to register is October 18th at 5PM. A virtual onboarding session is available here . 22:13 Matthew – “I think hackathons are fun. Good ways to learn things, good ways to get people interested. The only thing I question here is why are students not eligible?” AWS 24:22 AWS announces private IPv6 addressing for VPCs and subnets AWS is enabling Private IPv6 addressing for VPCs and Subnets within the VPC IPAM manager. On AWS, private Ipv6 addresses can take the form of unique local ipv6 unicast addresses (ULA) and global unicast addresses (GUA) and can only be used for private access. AWS does not advertise IPv6 addresses to the Internet. Within IPAM, customers can configure Ipv6 addresses in a private scope, provision ULA and GUA, and use them to create VPCs for private access. Customers use these Ipv6 addresses to boost security and assure compliance, as they can demonstrate that their resources with private IPv6 addresses are not internet accessible via a quick audit. 25:02 Matthew – “I love that they’re actually making IPv6 be simple to deploy, you know, the same way as the 10.8 and, you know, what is it, 192 & 168 and the other subnets that are private. I just don’t have a strong desire to deal with IPv6 nuances still in life. So I don’t foresee myself deploying this, but if you are a bleeding edge company, and or you want lots and lots of instances and or nick cards in the same subnet, you know, the same thing, go for it. It’s a great feature they’re adding.” 26:21 Amazon EFS now supports up to 30 GiB/s (a 50% increase) of read throughput Amazon EFS provides serverless, fully elastic file storage that makes it simple to set up and run file workloads with the AWS cloud. In March 2024, they increased the Elastic Throughput read throughput limit to 20 GiB/s from 10GiB, to support the growing demand for read-heavy workloads such as AI and ML. Now they are further increasing it to 30 GiB/s, extending EFS’s simple, fully elastic, and provisioning-free experience to support throughput-intensive AI and machine learning workloads for model training, inference, financial analytics, and genomic data analysis. 26:48 Matthew – “Better speed, always better. Faster speed, always better.” 27:19 Amazon CloudWatch Internet Monitor enhances dashboard and traffic suggestions Amazon Cloudwatch Internet Monitor has updated the console experience, including new features for visualizing configuration changes that can help you reduce latency for your application. With the refreshed dashboard, the internet monitor console now lets you easily find and take advantage of Internet Monitor’s breadth of capabilities. Want to visit the network monitoring page? Click here . 28:53 Announcing the general availability of AWS Backup logically air-gapped vault AWS Backup announces the GA of logically air-gapped vault, a new type of AWS Backup vault that allows secure sharing of backups across accounts and organizations. It also supports direct restore to help recover time from a data loss event. A logically air-gapped vault stores immutable backup copies that are locked by default and isolated with encryption using AWS-owned keys. You can get started with logically air-gapped vaults using the AWS backup console, API, CLI. Target backups to a logically air-gapped vault by specifying it as a copy destination in your backup plan. Share the vault for recovery or restore testing with other accounts using AWS Resource Access Manager (RAM). 30:07 Matthew – “I love that it’s actually managed for you end to end. I’m surprised that day one, I I looked, it wasn’t available in GovCloud because so many government restrictions require these things.” GCP 30:43 Query your data inCloud SQL Studio is GA for MySQL, PostgreSQL, and SQL Server | Google Cloud Blog seconds with Cloud SQL Studio Cloud SQL Studio for MySQL , PostgreSQL and SQL Server is now generally available. This gives you an in console, lightweight tool to query your database directly from the console. Cloud SQL gives you a consistent and intuitive user interface for all your databases regardless of engine. In addition to ease of access, you get the ability to quickly and easily create, edit and manage your database using an AI assistant that helps you write your queries with natural language. Cloud Sql Studio can help you take your database administration skills to the next level. 32:17 Justin – “Just doing = a quick little Google here, and people say you can, you can do it with things like Athena with like the JDBC drivers. That’s just not as clean in my opinion.” 34:58 Real-time in no time: Introducing BigQuery continuous queries for up-to-the-minute insights Data analytics and engineers are increasingly demanding expanded real-time capabilities to manage continuous data streams for both input and output. To address this challenge for customers, Google has transformed BigQuery into a real-time, event driven analytical platform, so they launched BigQuery continuous queries in preview. BigQuery continuous queries answers the challenge of cost and complexity of true real-time data analysis. Historically, “real-time” meant analyzing data that was minutes or even hours old. But with the demands for customer engagement, decision making and AI-Driven automation it’s now necessary to get this data in seconds. BigQuery continuous queries can execute SQL statements that can process, analyze and transform data as new events arrive in BigQuery, ensuring your insights are always up to date. The native integration with Google Cloud ecosystem unlocks even more potential. You can harness the power of Vertex AI and Gemini to perform ML inference on incoming data in real time, or if you want to replicate the results of continuous query to Pub/Sub topics , Bigtable instances, or even BigQuery tables for further processing and analysis. Continuous queries unlock several use cases: Simplify real-time pipelines: Express complex, real-time data transformations and analysis using the familiar language of SQL, removing the need for additional technologies or specialized programming skills. Unlock real-time AI use cases: Incorporate real-time data transformation with Google’s robust AI offerings using Vertex AI and Gemini, enabling a wide range of real-time AI-powered applications, such as generating personalized content, data enrichment and entity extraction, detecting anomalies instantly, and powering event-driven architectures. Streamline reverse ETL: BigQuery continuous queries integrate with other Google Cloud services like Pub/Sub and Bigtable, so you can send the results of a continuous query to Pub/Sub topics to craft event-driven data pipelines and Bigtable instances for real-time application serving. Alternatively, the results of a continuous query can be written into another BigQuery table for further analysis. Provide scalability and performance: Backed by BigQuery’s robust serverless infrastructure, continuous queries can handle massive volumes of data with high throughput and low latency. “At Bayer, we are under more pressure to deliver real-time analytics – which has historically proven difficult. Now that we’ve had an opportunity to evaluate BigQuery continuous queries, we are incredibly excited about the future possibilities this capability will unlock. From real-time integration of ERP, CRM, IOT data to real-time monitoring and alerting use-cases, we believe continuous queries will be a game-changer that will significantly expand the types of business challenges we can address within our data warehouse.” – Anthony Savio, Data Warehouse Engineering Lead, Bayer 37:37 Justin – “ I mean, it wasn’t like that into working with this type of thing as ksql, which is a Kafka SQL. And so basically as eventing kind of goes through that matches your query results through KSQL, you can pull it out immediately basically into tables and into other real time insights. So it makes sense that this would be something you’d want to build natively into BigQuery, especially considering the use cases that you have on that big data. So yeah, I’m glad to see this.” Azure 38:06 Announcing a new OpenAI feature for developers on Azure The latest open AI model is now available on Azure, but this one has the much less sexy name of GPT-4o-2024-08-06. Just rolls off the tongue, doesn’t it? This model brings innovative features, designed to elevate developer experiences on Azure. Specifically, the new model focuses on enhancing productivity through structured outputs, like JSON schemas, for the new GPT-4o and GPT-4o mini models. The system will provide two flavors of structured outputs User-defined JSON schema, where the developer specify the exact JSON schema they want the AI to follow More accurate Tool output (Strict mode) this limited version lets developers define specific function signatures for tool use, supported by all models that support function calling. 38:56 Justin – “I appreciate chat GPT being able to give me a structurally correct JSON schema that I’ve defined with my data set. That allows me to move it quickly to other systems that might need that data for input from JSON.” New Azure Data Box capabilities to accelerate your offline data migration Azure Data Box offline data transfer solution allows you to send petabytes of data into Azure Storage in a quick, inexpensive and reliable manner. The secure data transfer is accelerated by hardware transfer devices that enable offline data ingestion in Azure. Several new capabilities to Azure Data Box: GA of self-encrypted drives for Azure Data Box Disk SKU that enables fast transfers on Linux Systems. Support for data ingestion to multiple blob access tiers in a single order. Preview of cross-region data transfers for seamless data ingest from source country or region to select Azure destinations in a different country or region. Support in Azure Storage Mover for online catch-up data copy of any changes to active workloads that may have been generated post offline migration with Azure Data Box. Azure Data Box has also achieved HIPAA/BAA and PCI 3DS & PCI DSS certifications. 41:58 Matthew – “I do like here the data encryption of multiple blob access tiers in a single order. It’s been a long time since I used the Snowball. And I’ve always wanted to play with the Snow Cone. I just never did. But at one point, you can only dump it into EBS. And then from there, they added S3. And it was always like one account, like one location. And then especially when you’re moving data, especially if this is capable of doing up to petabytes.” 43:22 Unlocking the future of innovation: the Microsoft AI Tour Azure is setting out to go on Tour… an AI Tour of course. This is apparently the second year they’ve done this,and we apologize for not having caught this last year! The AI tour will visit 60 cities starting September 24th, offering a free one-day in-person experience on AI thought leadership. Sessions will help build AI skills, hands-on workshops will be offered, and connections will be made with other attendees who are practicing AI. We’re disappointed that they have not yet announced any dates on the west coast, But Matt can pick his poison – NY or Boston, maybe even international. He does love all things international.. Including pancakes. Matthew will be reporting back! Boston – October 1st 2024 NY – Jan 30th 2024 45:24 Public Preview: Customer Managed Planned Failover for Azure Storage In preview Azure is offering you a customer managed planned failover for Azure Storage is now available. Over the past few years, Azure storage has offered customer managed (unplanned) failover as a DR solution for geo redundant storage accounts. This allowed you to meet business requirements for DR testing and compliance. Planned failover now provides the same benefits, while introducing additional advantages. Planned Failover allows you to swap geo primary and secondary regions, while storage endpoints are still healthy. This allows you to swap between the two without data loss or additional costs. In addition, you will not need to reconfigure geo-redundant storage after the planned failover operation which will save you time and money. 46:03 Matthew – “I am excited for this, partially because I have to do a yearly. DR test. And this was something that last year we did, we had to have it copy all the data and then convert it back and then flip it back. it just hopefully gets rid of some of the monotony of the DR process.” OCI 48:46 Oracle Strengthens Saudi Arabia’s AI Economy with Opening of Second Public Cloud Region To meet the rapidly growing demand for Oracles AI and Cloud services , they are announcing their second region in Saudi Arabia. The new Riyadh cloud region will help public and private sector organizations migrate all types of workload to OCI. “The opening of Oracle’s new cloud region in Riyadh reflects the Kingdom’s continuous efforts in boosting the digital economy based on modern technologies and innovation,” said His Excellency Eng. Haytham Alohali, vice minister, the Ministry of Communications and Information Technology. “This step will significantly enable international and local companies to achieve innovation and promote the adoption of AI and cloud computing technologies in various sectors, which enhances Saudi Arabia’s competitiveness at the regional and international level.” “With the rapid expansion of our cloud footprint in Saudi Arabia, Oracle is committed to helping the country achieve its goal of developing one of the strongest digital economies in the world,” said Richard Smith, executive vice president and general manager, EMEA Cloud Infrastructure, Oracle. “As part of our wider investment in cloud capabilities in Saudi Arabia, the Oracle Cloud Riyadh Region will help accelerate adoption of cloud and AI technologies to boost innovation across all sectors of the Saudi economy, while helping organizations address local data hosting requirements.” 49:40 Matthew – “They’ve shipped another truck to Saudi Arabia.” Closing And that is the week in the cloud! Visit our website, the home of the Cloud Pod where you can join our newsletter, slack team, send feedback or ask questions at theCloud Pod.net or tweet at us with hashtag #theCloudPod