I’ll admit that I watched the 2018 Nathan’s Hot Dog-Eating Contest. Don’t judge me. It’s awe-inspiring—and super gross—to see a human being consume 74 hot dogs in ten minutes. But it’s an annual tradition, and highlights a food item that’s associated with summertime. We often adjust our diet based on the season or what we’re into at the moment. There’s no doubt that you’re into public cloud right now, and making it a big part of your I.T. “diet.” I don’t think public cloud is a seasonal fad—it’s representative of a long-term change in how you consume infrastructure.
Just as you need to prepare yourself to ingest dozens of hot dogs in a few minutes, you need to ask some important questions before consuming lots of public cloud. Is lock-in a real thing? What if you’re using more than one cloud? How do you decide what goes where? Most of Pivotal’s customers are multi-cloud at this point—heck, we just won Microsoft’s U.S. Azure Consumption Partner of the Year award again—so we have a unique perspective that may help you answer those questions.
Let’s look at three things we’ve learned over the past year.
You won’t be “all in” with a single cloud, but not for the reasons you think
Infrastructure-as-a-Service is growing like crazy. Gartner pegs that growth at 29.5%. What’s fascinating is that that growth isn’t tied to a single vendor. And, on-premises server sales continue to rise as well. 451 Research discovered that a significant number of companies prefer to run apps on multiple infrastructure pools, and are regularly moving workloads between public and private environments. Their reasons for using multiple clouds? Survey respondents cited improving performance and availability, cost optimization, data isolation, and extending existing investments. Smart organizations choose a multi-cloud strategy because they want control when they need it, access to best-of-breed services, a global presence, and smart utilization of current assets.
In recent months, Pivotal VP Josh McKenty and Microsoft Distinguished Engineer Ulrich Homann hit the road together, talking to customers to hear about their challenges first hand. What did they hear? Customers are excited about the economics and agility benefits of public cloud. They’re jazzed about the developer and operator gains of cloud-native platforms like Pivotal Cloud Foundry (PCF), but are hesitant to single-source and risk cloud lock-in.
Everything has lock-in, as I’ve written about before. And CIOs are consistent in their view about this: over-rotating to avoid cloud lock-in by developing in a cloud-agnostic fashion limits the benefits they want from public cloud. Your task is to find the balance between innovation, flexibility, and portability. In a recent webinar, we offer guidance to help you do just that.
Josh and Uli shared their roadtrip insights and perspective on multi-cloud, debunk many of the industry myths around cloud portability:
https://content.pivotal.io/webinars/may-31-debunking-myths-about-cloud-portability-webinar
Figure out where you’re ok using proprietary services, and how to still get a level of portability
All clouds aren’t created equal. While there’s a commoditization happening at the infrastructure layer—think virtual machines, storage, networking—each cloud provider offers unique value. For example, in Microsoft Azure, you have a variety of novel services with heavy investment: databases, machine learning, application integration, and their overall “intelligent cloud” model. How do you decide when to use commodity capabilities versus coupling to cloud-specific services?
Uli and Josh discuss this in the above webinar. Specifically, to use innovative tech like artificial intelligence or function-as-a-service, you might not have a commodity option! Another scenario? For fast moving experiments, you’re often focused on quick learning, and coupling to proprietary services is reasonable at that stage of your software’s lifecycle.
But your multi-cloud strategy is going to require some level of shrewd standardization. If you use every cloud’s proprietary services—even for fairly standard things—you’re going to massively increase your cost and complexity. But how do you get a level of consistency across clouds?
-
Look for “Uli’s Separation” to get standard interfaces on novel services. In the webinar mentioned above, Microsoft’s Uli Homann talked about a strategy for using per-cloud services, but in portable ways. He called out two examples. First, Azure’s CosmosDB database is a powerful, proprietary NoSQL database. But, it also has APIs that mimic both MongoDB and Cassandra databases. This way, you can use CosmoDB, but retain a level of portability in your code. Similarly, the Azure Event Hubs service is only found in Microsoft’s cloud. However, it has a new Apache Kafka interface, which lets you bring apps, unchanged, to this service. Keep an eye out for unique cloud services that serve up standard interfaces.
-
Use platforms that automate the underlying infrastructure and standardize operations. One proven way to achieve multi-cloud effectiveness is to use a platform that runs atop every cloud. This gives your developers the option to run anywhere, while avoiding bespoke automation for each provider. Pivotal Cloud Foundry doesn’t just automate the installation, upgrade, and management of cloud IaaS infrastructure. It also introduces a uniform way to collect application and infrastructure logs, analyze infrastructure metrics, manage credentials, deploy software, and isolate tenants.
-
Find the right number of options for developers. As Pivotal’s Josh McKenty will tell you, developers love choice. But it’s often about providing just enough choice to avoid unnecessary constraints, while limiting the endless options that fluster your operations team. Define a handful of programming languages, database engines, and messaging technologies for your team to choose from.
-
Leverage service brokers to access per-cloud services. One of your most important takeaways from this story? Creating a consistent experience across clouds does not mean you can’t hungrily consume the unique services of each cloud. But do it in a consistent way. Enter service brokers. The Open Service Broker API originated in Cloud Foundry and is now in use across platforms. What this does is define a standard way to advertise services and bind your apps to them. Microsoft maintains their broker and they make it easy to use a wide range of unique Azure services from within Cloud Foundry or Kubernetes environments. And don’t forget, you can use any service broker in any cloud. Use the Azure broker even from your on-premises data center!
For more hybrid stats and strategies, check out this free Forrester report commissioned by Pivotal and Microsoft.
Partner with those who “get it”
Crafting and executing a multi-cloud strategy isn’t easy. Modernizing your existing software to take advantage of this play isn’t any easier. Who out there really knows how do it? Pivotal does. And Pivotal and Microsoft have proven to be remarkable partners that unify on-premises and public cloud environments for the biggest brands in the world. Let’s look at what we’re doing together to make your multi-cloud journey a success.
https://twitter.com/ncayou/status/1018686173202415616
When you work with Microsoft and Pivotal, you get the result of powerful joint engineering. One example? Get access to a regularly updated BOSH CPI for Azure (and Azure Stack) owned by the Microsoft team. Or use the powerful Azure service broker for connectivity to brilliant services. Operators have an integrated management viewpoint thanks to PCF integration with Azure’s Operations Management Suite. Developers get a host of Spring Boot integration, including Spring Boot Starters for Azure services and Visual Studio Code extensions. And don’t forget about Steeltoe which makes it simpler to deliver cloud-native .NET microservices. Use Azure Active Directory for SSO within PCF. And get fully supported Windows Containers wherever you run PCF.
You need more than just great tech. You need assurances that you can resolve issues quickly, and constantly get new value from this partnership. Microsoft and Pivotal have cross-trained support teams, data sharing agreements, and joint phone bridges to quickly collaborate on customer problems. On top of that, our two organizations meet regularly to share roadmaps and look for places to add new value for you.
Finally, you’ll see Pivotal and Microsoft continue to partner in the community to share best practices and training for multi-cloud scenarios. We’re both part of the Cloud Foundry Foundation. Pivotal joined the .NET Foundation and donated Steeltoe to it. Both companies are members of the Cloud Native Computing Foundation. Microsoft is active in Pivotal’s Spring community and we’ve partnered on a series of hands-on Java workshops. Pivotal’s working with key Microsoft partners like Magenic to deliver cloud-native .NET workshops. Pivotal and Microsoft have traveled the world together to talk to you, and expect to do even more in the year ahead.
Come hang out with Pivotal and Microsoft at SpringOne Platform, and look for a workshop in a city near you!
