Google Cloud Next 2018: What GCP’s Customer Focus Means for You
Google Cloud Platform (GCP) has been steadily gaining momentum as one of the three vendors in the Gartner IaaS Magic Quadrant. And it shows. Google Cloud Next ’18 was the largest one yet with 25,000 registrants, hundreds of speakers and a rocking Chainsmokers concert.
With keynote titles like “Building a Cloud for Everyone,” “Bringing the Cloud to You” and “Made Here Together,” it’s no surprise that the 105 announcements were focused on enabling customers, especially enterprise customers, to better use Google’s cloud building blocks to achieve their goals. Of those 105 announcements, there were a few that really stood out to us.
We’re seeing more and more companies adopting multi-cloud best of breed architecture. We think these features are going to play a big part in how GCP is included in that mix.
Cloud Services Platform Brings Modern Building Blocks to Enterprises
Google’s leadership in containers is spearheaded by their open source Kubernetes project. When we built our State of the Cloud 2018, we saw ridiculous growth of Kubernetes as the container management platform of choice (from 40% to 70% market share). We expected more growth in 2018, but that growth is now poised to explode with Google’s new Cloud Services Platform. Cloud Services Platform is actually a whole family of tools companies can use to manage their cloud services.
What really caught our interest with Cloud Services Platform is how it enables the adoption of cloud-based technologies for cloud and on-prem systems with the same framework. Time and time again, the lack of easily portable application frameworks and infrastructure has been a major cloud migration hurdle for our customers. Integrating containers and microservices into that migration adds additional challenges that are magnified exponentially when you’re working at an enterprise level. Cloud Services Platform makes it easier.
Organizations now have the flexibility to develop applications using containers and microservices on-prem, then migrate to the cloud as they increasingly get comfortable with them. It also means that hybrid cloud situations can be managed using the same suite of services, making it easier to manage cloud technologies on-prem and making data centers a stronger stepping stone to the cloud.
As big believers in helping customers do more with cloud, this made us incredibly excited. We’re going to dig into a couple of those features and let you know how we think they’ll impact the cloud — especially cloud costs.
Last year, Google kicked off Istio as an open-source project to deliver a service mesh that enables microservices architecture. If you’ve been waiting for that work to be realized, then your wait is over: Istio v1.0 has been released.
Developing and maintaining microservices has always required a multi-faceted approach that tackles challenges like service discovery, load balancing, networking and security capabilities. The Istio platform is designed around three core capabilities to address those challenges:
- Security: All microservices communication is inherently secure via encryption and AuthN and AuthZ. Defense in depth security and security policies are delivered at both the application layer and the network layer.
- Monitoring: Service visibility via logging and monitoring capabilities delivers visibility for development and operational teams. We’ll get into this more below when we talk about Stackdriver.
- Control: Policy-driven controls help organizations control traffic using policies specific to their application and environment.
Istio is also available in the Kubernetes Engine, which is going to spur more microservice and container architecture use. We think the ease of microservice management is going to let companies get more out of their cloud. We also anticipate that increased adoption creating cloud cost management challenges. Since those are exactly the challenges we help solve, we’ll be watching the market very closely now that Istio has been released.
Stackdriver Monitoring is a brilliant compliment to Istio’s service mesh, and its integration takes monitoring to the next level. As more applications are increasingly run on ephemeral containers with microservices and auto-scaling capabilities, it’s getting tougher to get detailed performance and operational visibility into your services. Stackdriver helps provide that visibility. Customized alerts let you instantly know when predefined limits are violated, while dashboards let you dig in deep to find the issue.
Providing visibility into complex services is always a tricky problem, and the distributed nature of microservices make things extremely complex. The Stackdriver tool is tackling the problem by striving to provide a simple application view, which we think is going to be very useful for developers and operators. We wouldn’t be surprised if this increased visibility was a major factor in the greater adoption of microservices.
Google Cloud continues to establish Kubernetes as the de facto container solution with the announcement of GKE On-Prem. Currently accepting applications for alpha early access, GKE On-Prem allows companies to deploy a Google Kubernetes environment to their on-prem environment. Best of all, the solution securely connects with Google Cloud management console, giving users a single pane of glass for all of their containers — whether on the cloud or on-prem. The demos at Next were very impressive and showed off how seamless the experience will be.
We’re really excited to see how this feature is used by enterprise companies. It promises an easier way for organizations new to the cloud to familiarize themselves with with containers and a microservices architecture. And once they take the leap to the cloud, they won’t have to learn a whole new management system. At the same time, it gives development teams the chance to make sure their applications are “cloud ready” before deployment.
This has the potential to be a major boon for enterprises. We think there are plenty of companies that will embrace the opportunity to seamlessly develop, operationalize and manage across on-prem and Google Cloud. We also love the idea of a unified single pane of glass and have developed our platform with that tenant in mind.
Cloud Functions and Serverless Containers
Google’s Cloud Functions is now generally available, kicking GCP’s serverless capabilities into high gear. Along with general availability, Google also threw on a few new features like support for Python 3.7 and Node.js 8, networking and security controls, and their SLA predictable service guarantee.
But we’re not going to lie: we got a little distracted by their announcement of serverless containers. Using serverless environments has always involved a tradeoff. The managed runtime environment means you don’t have to worry about the infrastructure, but it also means that you don’t have as much runtime flexibility. Google’s innovative solution combines the flexibility of containers with the managed infrastructure of serverless environments. Just think about it! Say you have code and libraries currently running as a container which are not natively supported in serverless. Now you can run it as as serverless container in Google Cloud. You’ll get the benefits of containers while still only paying for what you use.
Serverless containers are currently only available via early access for Cloud Functions or as the GKE serverless add-on. Regardless, we’re going to be watching this space closely. Serverless containers could have a big effect on the evolution of microservice architecture over the next year, and, in turn, how companies use the cloud. We’re especially excited about this development having seen the exponential growth of container and serverless in our customer cohort. As customers increasingly adopt new technologies, our platform delivers the ability to allocate costs for them.
Knative is an interesting open-source initiative that was launched in partnership with Pivotal, IBM, RedHat and SAP. Built on the same technology behind the GKE serverless add-on, Knative is designed to give you the same building blocks needed to build and deploy serverless containers without being tied specifically to Google Cloud.
This cloud-agnostic approach grabbed our attention, particularly since we’re a multi-cloud platform. We applaud Google for taking the initiative, and think it’s going to drive forward their continued industry leadership in the container space.
All in all, Cloud Services Platform is a big step for containers and Kubernetes. We can’t wait to see how it shakes up the cloud space!
Sustained Use Discounts Get Enhanced
Beyond the world of Kubernetes, we were also very excited to see an upgrade in how Google will apply sustained use discounts for Google Compute Engine. Sustained use discounts work by giving you a discounted rate based on your use. The more sustained use, the more discount you get. Before now, the discount was applied based on a particular machine type, which means you might miss out if you changed your machine type too often. But not anymore. Google now applies discounts holistically based on the vCPU and memory usage in a region, regardless of machine type. This is a great step forward which means customers get discounts for their aggregate usage and greater discounts overall. And we always like to see cloud providers making it easier for customers to get more from their cloud spend.
That being said, getting the most out of sustained discounts can still require some strategic planning. (For more, check out our GCP best practices e-book.) The sustained discount only applies for the resources used within the same region. Also, customers are going to save much more by taking advantage of committed use discounts and finding that perfect blend. Fortunately, we believe that data science and machine learning can help companies get the most out of their cloud. As our founder and CEO Mat Ellis likes to say, “We aim to dip every feature our customers ask for in a cup of data first before releasing it into the wild.”
New Capabilities Democratize AI
Google’s tensor processing units (TPUs) were on full display at Next, and there was plenty to geek out about. The Cloud TPU units on display went from TPU v1 all the way to the water-cooled TPU v3, which can process 11.5 petaflops! For all the builders out there, it was a sight to behold. Google continues to build on its AI technology lead by announcing several key new capabilities:
- BigQuery ML: Many organizations lack in-house data science capabilities. BigQuery ML is designed to help. Since a lot of data is already stored in BigQuery, companies will be able to create and deploy machine learning models faster.
- Kubeflow v0.2: Kubeflow was announced last year as an open-source solution to leverage machine learning stacks such as TensorFlow and Scikit-Learn on Kubernetes. Kubeflow v0.2 features an improved UI and enhancements to monitoring and reporting.
- Second and Third Gen TPUs: Want to take advantage of those fancy TPUs? Second gen TPUs are now generally available, and the alpha release of Cloud TPU Pod can tap into that 11.5 petaflop processing ability. At the same time, Google also announced the new Edge TPU chip and Cloud IoT Edge.
More and more companies are adopting machine learning and AI — more than 60% by some counts. Most of the power behind that is going to come from the cloud, and Google is making sure they’re ready. Our State of the Cloud 2018 analysis showed customers adopting GPUs at a rapid rate, so we expect companies to leverage machine learning capabilities into a competitive advantage and disrupt industries.
Google’s Working Hard to Stay Ahead of Their Customers
Google Cloud Next 2018 was an exciting show for Cloudability. Google doubled-down on their commitment to open source, containers, microservices, AI and generally enabling their customers to push the boundaries of what’s possible with the cloud. By partnering with some of the biggest names in IT like Cisco, SAP, Red Hat and Pivotal, GCP is creating the building blocks for organizations to build comprehensive hybrid cloud systems around their needs. All in all, Google Cloud demonstrated that the only effective solutions are ones that are built around the customer.
We couldn’t agree more. Cloudability always focuses on doing what’s best for customers and on helping our customers solve their problems. Having recently rolled out support of GCP on our cloud cost management platform, we can’t wait to see what our customers will do with Google Cloud.
Sign up for your free trial and find out how you can get more from your cloud spend.