213: The Cloud Pod Sings a Duet About AI

Welcome to the newest episode of The Cloud Pod podcast! Justin, Ryan, Jonathan, Matthew are your hosts this week. Join us as we discuss all things cloud, AI, the upcoming Google AI Conference, AWS Console, and Duet AI for Google cloud. 

Titles we almost went with this week:

• You can finally lock yourself out of the AWS Console! 
• Google IO delivers the AI… hopefully soon to be renamed Google AI Conference
• ️Azure announces major MySQL upgrade!
• Azure can now update mysql without taking itself offline

A big thanks to this week’s sponsor:

Foghorn Consulting, provides top-notch cloud and DevOps engineers to the world’s most innovative companies. Initiatives stalled because you have trouble hiring?  Foghorn can be burning down your DevOps and Cloud backlogs as soon as next week.

News this Week:

01:10 - Terraform is in the news! 

• Terraform Cloud updates plans with an enhanced Free tier and more flexibility
• A bunch of new updates are coming to Terraform Cloud
• These update will provide access to **premium** features, up to 500 resources in the free tier
• There are also new paid offerings for management capabilities, scaling currency, and enterprise support. 
• Consistent billing metrics based on managed resources, scaling concurrency, and enterprise support area available across all tiers. But let’s be honest - who needs consistent billing metrics? Half the fun is in the guessing!
• New Features Include:
  • Premium security features such as SSO and Policy as Code on all tiers (yes, even the free ones for the poors like us.)
  • Make it “easy and frictionless” for smaller teams and organizations to get started with their first use cases. 
  • And -finally- updated paid tiers provide easy upgrade paths for organizations as their usage scales, and they have more advanced use cases.
• Consumer Advice Time! The updated pricing models include a “per resource” charge. That has the potential to get REAL messy over 500 devices. 
• Of course, it’s an option to stay on the legacy models, but the “carrots” - like SSO and Sentinel/OPA support - are pretty good, so you really just need to do a cost benefit analysis for your particular situation. 

02:35 Ryan - “Yeah, I mean, the licensing for Terraform products for cloud and both enterprises always been rough, right? Like starting off per users for cloud makes sense. And at some point for enterprise, they had switched to per project, not users, because they figured out very quickly that what everyone did was just sort of link it together behind automation pane.” 04:48 ”Justin - the devil's in the details of what they consider a resource, right? And it's every single thing. I mean, 10 cents per EC2 instance, hmm. Like, yeah, I get 10 cents worth of value out of Terraform, not having to manually do that stuff. So, like, yeah, but then like you get into  S3 buckets and like, I'm definitely not gonna get 10 cents of value out of an S3 bucket every month.”

• Our only big question from this announcement is just what they consider a billable resource vs. a supportive resource. Example pricing could potentially be really helpful here. 
• Does anyone in the audience have a PhD in finance? We could use some help with some cost calculation. 

AWS

11:00 Amazon Aurora I/O-Optimized Cluster Configuration with Up to 40% Cost Savings for I/O-Intensive Applications 

• **Announcing!** The general availability of Amazon Aurora I/O Optimized!
• This new cluster configuration offers improved price performance and predictable pricing for customers with I/O intensive apps (like e-commerce, payment processing systems, pretty much anything with SAP)
• As much as we like the “everybody loves a surprise bill” method, you can now more confidently predict your costs for I/O intensive workloads - and get up to 40% cost savings when your I/O exceeds 25 percent for your current Aurora Spend. 
• If you use reserved instances for Aurora, you could potentially see even greater savings. Rad! 

13:10 Jonathan - “The predictability of the workload means that they can, that Amazon themselves can better kind of put customers in buckets for IOPS. And so they can manage capacity better, whereas customers with very bursty workloads, they always have to make sure that capacity's available when they need it.” 14:13 Private Access to the AWS Management Console is generally available 

• Now generally available in AWS Management Console - Private Access!
• Private access allows access to your AWS Management Console from on-premise networks using a secure, private connection. 
• AWS touts the new access as easy to set up, and a good way to help improve costs, security, and compliance. And who doesn’t love improvements to cost, security, and compliance? 
• Some of the benefits of using Private Access to the AWS Management Console:
  • Improved security: Private Access to the AWS Management Console provides a more secure way to access AWS resources, instead of using a public internet connection.
  • Increased compliance: Private Access to the AWS Management Console can help you meet compliance requirements by providing a secure, controlled way to access AWS resources.
  • Reduced costs: Private Access to the AWS Management Console can reduce costs by eliminating the need for a VPN or other expensive connectivity solutions.
  • Improved performance: Private Access to the AWS Management Console can improve performance by providing a direct, private connection to AWS resources.
  • Increased flexibility: Private Access to the AWS Management Console can provide increased flexibility by allowing you to access AWS resources from your on-premises network.
• Want to start utilizing Private Access? Us too! It is currently available in US East (Ohio), US East (N. Virginia), US West (Oregon), Europe (Ireland), and Asia Pacific (Singapore).
• Looking for a more secure, compliant, and cost effective way to access your AWS resources? Private Access might just be the answer. Allegedly. 

**insert a classic Matthew horror story about some weird tech niche** 17:14 Using Open Source Cedar to Write and Enforce Custom Authorization Policies

• Amazon has released support for Cedar 
• Cedar provides a simple and intuitive API that makes it easy to write policies for all your applications.
• Cedar supports a variety of authorization models, including role-based access control (RBAC), attribute-based access control (ABAC), and capability-based access control (CBAC).
• Cedar is highly extensible, allowing you to customize it to meet your specific needs.
• Cedar is well-documented and has an active community of users and contributors. 
• Cedar has a pretty website. 

17:52 Matt - “This is the first I'm hearing of this kind of concept and I'm loving it. Cause this is one of those things where if you're building your own app, I dread sort of the authentication flow, and so I'm always trying to leverage some other party for this.” 18:26 Justin - “The examples they gave, this tiny to-do user policy thing, it's pretty great. It's a cute little app, and it gives you a very simple way to use it in a client, like a Python client or a Rust server. It's not a bad little example of how to do it. So yeah, I love all this. Ever since CDK, everybody's rushing to make all these things as code, but not restricted by CloudFormation. So I'll take it as a win anytime, it's not CloudFormation.” 19:14 Announcing Provisioned Concurrency for Amazon SageMaker Serverless Inference

• More exciting news from Amazon! SageMaker Inference now offers provisioned concurrency. This will allow you to set a specific number of concurrent requests that your model can handle.
• This can help you to ensure that your model is always available to serve requests, even during periods of high demand.
• Provisioned concurrency is available for both batch and real-time inference.
• To use provisioned concurrency, you need to create a SageMaker endpoint with the "provisioned" concurrency mode.You can then specify the number of concurrent requests that you want your endpoint to handle.
• What are the costs for this new Provisioned Concurrency for Amazon SageMaker Serverless Inference magic we hear you asking? Well, dear listener, wonder no more. 
  • Pricing is based on the number of concurrent requests that you provision, and if you choose to provision one or multiple models. 
  • The cost for provisioned concurrency is per hour, and you are billed for the hours that your concurrency is provisioned. For example, if you provision 10 concurrent requests for a model that costs $0.01 per hour, you will be billed $0.10 per hour for the provisioned concurrency. If you only use 5 concurrent requests for that model, you will only be billed $0.05 per hour. 
• Savings Plan is available for provisioned concurrency; and can offer a lower per-hour price than the on-demand price. 
• Users can choose between 1 or 3 year Savings Plan. 
• The bulk buying rule applies here - the more concurrency you commit to, the lower your price per hour will be. 

20:08 Jonathan - “Seems to be becoming quite a pattern, doesn't it? They've moved away from servers, have serverless, it's on demand, you pay for what you use, but also now you can have provisioning capacity because you realize that that didn't actually work for people.” 20:08 Ryan - “Well, it's not that it didn't work. It's that the people having to pay for this do not like spiky, unpredictable workbooks. Right? Like a lot of the provision capacity isn't because you run out of capacity. It's for consistency. And so like if you're going to use this, use it, right. And then we'll baseline the provision part so that we have a consistent cost model. Because otherwise we have no idea what our costs are doing. We don't know when it's out of control or we don't know when it's normal. And it's, so a lot of this I think is, is less about actually having the capacity to execute than it is actually just standardizing and removing those big spikes.”

GCP

22:56 At Google I/O, generative AI gets to work 

• Lots of news coming out of Google I/O this year, and surprise surprise - most of it centers around AI! Who would have guessed? Not us…
• According to Google CEO, easy and scalable AI is going to drive innovation at all levels of business. 
• Google Cloud announced a number of new AI and machine learning products and features at Google I/O 2023. These include:
  • Cloud AutoML Natural Language, which makes it easier to build natural language processing models without any machine learning expertise. (Justin is excited for this one. Natural language for the win.) 
  • Cloud AutoML Vision Edge, which enables developers to build and deploy custom vision models on edge devices.
  • Cloud TPUv4 Pods, which are powerful machine learning accelerators that can be used to train larger and more complex models.
  • Cloud ML Engine Pipelines, which make it easier to manage and monitor machine learning pipelines.
• Cloud AI Platform, which is a unified platform for building, training, and deploying machine learning models.
• These new products and features are designed to make it easier for developers and businesses to build and deploy AI and machine learning solutions. They also “demonstrate Google's commitment to AI and machine learning, which are two of the most important technologies of our time.”
• And that’s a recap on Google AI, er, I mean Google I/O for 2023. 

26:15 Introducing Duet AI for Google Cloud – an AI-powered collaborator 

• A few weeks ago we discussed BARD, and how Google had added some new programming languages and capabilities. 
• At Google I/O they have now presented Duet AI for Google Cloud.
• Duet AI, which is on a limited access (much to Jonathan’s dismay) will help create a cloud experience that’s more personalized and intent-driven.
• Duet AI will understand your environments, and assist users in building secure and scalable applications - all with expert guidance. 
• New capabilities powered by Duet AI
  • Code Assistance which provides AI-driven code assistance for cloud users such as application developers and data engineers.  It provides code recommendations as they type in real time, generates full functions and code blocks, and identifies vulnerabilities and errors in the code, while suggesting fixes. 
  • Chat assistance to get answers on specific development or cloud-related questions. Users can engage with chat assistance to get real-time guidance on various topics, such as how to use certain cloud services or functions, or get detailed implementation plans for their cloud projects. 
  • Duet AI for Appsheet will help you create intelligent business applications, connect their data and build workflows into Google Workspace via natural language. 

28:28 Justin - “So one of the things they announced at Google I.O., that I don't have a story here for us, but they announced integration of BARD and stuff into Google Apps, and you could subscribe for the beta. And so I got the CloudPod Google Workspace into the beta. So most of the show notes today were written by AI.” (**Insert tears from the show note writer.**)  ☎️Listener poll: What sorts of interesting, non-WGA line busting uses does Chat GT have in your day job (beyond dad jokes and basic chat)? Let us know! 

Azure

34:21 New and upcoming capabilities with Elastic Cloud (Elasticsearch)—An Azure Native ISV Service 

• Moving on to Azure and services we hate with Elastic Cloud (missed title opportunity.)
• Elastic Cloud Elasticsearch is now an Azure native ISV service. It’s now fully integrated with Azure services and can infect, I mean manage, using Azure tools.
• There are a number of new and upcoming capabilities in Elastic Cloud Elasticsearch, including:
  • Support for Azure Kubernetes Service (AKS)
  • Integration with Azure Synapse Analytics
  • Support for Azure Active Directory (Azure AD) authentication
  • Improved performance and scalability

36:33 Jonathan - “You know, when people are addicted to drugs, we don't just say well here's a drug dealer that's got a better deal we say maybe you should move off that and try doing something else.” And that pretty much sums up the guys’ thoughts on Elastic Cloud. Moving on.  37:25 What’s new with Azure Files 

• Azure Files is a fully managed file storage service that provides SMB and NFS file shares. It is a highly available, scalable, and durable service that can be used to store any type of data, including application data, user data, and backups.
• There is currently no premium tier, but let's be real - it’s coming. Just wait and see. 
• Azure Files offers a number of features, including:
  • Durability: Azure Files is designed to be highly durable. Data is replicated across multiple data centers to protect against data loss.
  • Scalability: Azure Files can be easily scaled up or down to meet your needs.
  • Performance: Azure Files offers high performance for both reads and writes.
  • Availability: Azure Files is available in all Azure regions.
  • Security: Azure Files offers a number of security features, including encryption and access control.
• New features include
  • Azure Files now supports Azure Files Sync, which allows you to sync files between Azure Files and your on-premises file servers.
  • Azure Files now supports Azure Files Premium, which offers higher performance and scalability.
  • Azure Files now supports Azure Files Edge, which allows you to deploy Azure Files to your edge locations.
• The cost of Azure Files will depend on the following factors:

• The type of storage account you choose.
• The amount of storage you need.
• The performance level you need.
• The region you choose.
• For example, a Standard storage account with 1 TB of storage in the US East region would cost \$12.50 per month.
• Just know this doesn’t include network data transfer costs - which will most likely be *high*. 

40:10 Ryan  - “The Cloud. Full of sharp edges.” (Yet another missed title opportunity.) 40:41 General Availability: Azure Database for MySQL - Flexible Server major version upgrade 

• Azure Database for MySQL Flexible Server now supports major version upgrades.
• This means that you can now upgrade your database from one major version to another without having to rebuild your database - THANK GOODNESS! 
• The upgrade process is fully automated and takes care of all the necessary steps, including data migration and schema changes.
• To use the major version upgrader, you must have a subscription to Azure Database for MySQL Flexible Server.
• You can also find more information about the upgrade in the Azure docs if you want to learn more about this. 
  • We don’t. 

41:22 Jonathan - “I mean, if you can roll back, that's nice. If there's less downtime, that's nice. If it syncs ahead of time before it does it, that'd be kind of cool. But saying that it does all the necessary steps? Yeah, I don't think so. There's a whole lot of testing involved in major upgrades for MySQL.” 41:42 Ryan - “ALL THE UPGRADES. It said ALL the upgrades! I’m just gonna click the button.” 41:48 Matt - “What could possibly go wrong?”

Oracle

No new news. Sad face. 

Continuing our Cloud Journey Series Talks

47:27 Managed Services

• Welcome back to Cloud Journeys! We’re still talking cloud native, and this week we’re focusing on managed services. In cloud native, there are 5 things managed services should be providing you with - and those include: 

1. Reduces costs. Managed services can help you reduce costs by offloading the responsibility of managing and maintaining infrastructure to the cloud provider. This can free up your team to focus on other tasks, such as developing and deploying applications.
2. Increases agility. Managed services can help you increase agility by providing you with access to the latest technologies and features. This can help you quickly develop and deploy new applications and services.
3. Improves reliability. (Unless you’re in France. Too soon?) Managed services can help you improve reliability by providing you with a high level of uptime and availability. This can help you ensure that your applications and services are always available to your users.
4. Reduces complexity. Managed services can help you reduce complexity by providing you with a single point of contact for all of your cloud needs. This can help you avoid the hassle of managing multiple vendors and platforms.
5. Improves security. Managed services can help you improve security by providing you with a secure environment for your applications and data. This can help you protect your business from cyberattacks.

47:58 Jonathan - “Reduce cost is interesting because I think they can enable better architectures, thinking about serverless and event driven architectures which don't cost anything. That could reduce costs versus running something 24-7. However, managed services in general are really not cheaper than running them yourself. It's just a matter of where you spend the money, I think.” 48:24 Justin - “Well, you have to calculate the ROI differently. Like an RDS database, for example, you know, yes, you're offloading SQL Server Management or MySQL Management or Postgres or whatever flavor of database you're using to the cloud provider. So maybe you have 10 DBAs, and now maybe you only need six DBAs. So you have an ROI there, because you were able to do less DBAs, or have the DBAs do a more valuable thing, you don't have to fire them necessarily… And if you can focus on just your app, then you save money. But that ROI is not a direct ROI because that database costs 20% more than that database would have cost you on EC2. But you have less headcount required to support it.” ☎️Listener Poll - what sorts of things can you think of that are advertised as a managed service but it really isn’t? Matt’s example - having to tell Azure the number of servers that run your load balancers.  55:33 Ryan - “That's my biggest gripe with Composer; is that it's pretending to be a managed Airflow service and it's not. It's a deployment template.” Keep listening for some after show convos with Justin and the boys - especially if you’re interested in learning about when Yahoo was cool. (It was! It really was!)

After Show

1:00:01 Can Marissa Mayer Eclipse Herself? 

• Marissa Mayer was the CEO of Yahoo! from 2012 to 2017.
• She is now the CEO of Lumi Labs, a company that develops augmented reality technology.
• The article discusses whether Mayer can "eclipse herself" at Lumi Labs, as she did at Yahoo!.
• The article argues that Mayer's success at Yahoo! was due to a number of factors, including her experience at Google, her strong leadership skills, and her ability to attract top talent.
• The article also argues that Mayer's success at Lumi Labs will depend on a number of factors, including the company's ability to develop successful products, the market for augmented reality technology, and Mayer's ability to lead the company.
• The article concludes that it is too early to say whether Mayer will be successful at Lumi Labs, but that she has the potential to be a successful entrepreneur.

1:01:56 Ryan - What I've been waiting for from Lumi Labs is sort of like… give me something to play with because their application availability and stuff is very early and not generally available. I kind of want to see what they'll do. I think that one of the things that she spearheaded while I was at Yahoo was a lot of the weather app stuff and her views on mobile - for a company at the time that was really trying to figure out whether it was a content company or a technology company - her views on mobile were very different from what we were used to for the last few years since the previous CEOs. She was very opinionated, very data driven, and had introduced some really cool mobile experiences during that time. So I think that she's got a good track record of that kind of delivery. So I look forward to what they're doing.”

Spotted on the Horizon

Next week on the Cloud Pod Podcast…

Closing

And that is another week in the cloud! We would like to thank our sponsors Foghorn Consulting, who is definitely NOT AI generated. Check out our website, the home of The Cloud Pod where you can join our newsletter, slack team, send feedback or ask questions at thecloudpod.net or tweet at us with hashtag #thecloudpod