This week Google Cloud Platform held its yearly conference, Google Cloud Next. The event illustrated what to expect going forward and the company’s growth. Diane Greene, CEO of Google Cloud, emphasized their goal of “playing the long game.” Since GCP tends to be looked at as an outside, this approach makes sense. The company offers a variety of unique services and tools, meaning they will inevitably catch up. They put the power of Google’s internal developments in the hands of enterprises, which can make for a potent combination!
Managed Service Providers and Oracle Workloads
Cloud partners can indicate the popularity of a cloud platform. AWS has an abundance of partners; so many that an entire market has formed under the AWS umbrella. GCP trails behind both Azure and AWS when it comes to partners. GCP was the last of these clouds to release commercially and they’ve been working to catch up ever since. Thus, managed services were a topic of discussion at Google Cloud Next.
GCP announced MSPs can provide services for Oracle workloads. The services can be fully managed by partners. These partners work with users to leverage existing licenses and manage workloads. The services will receive full support from GCP.
Intel and SAP Partnership
This week, GCP announced a partnership with Intel and SAP to offer GCP virtual machines that support the upcoming Intel Optane DC Persistent Memory for SAP HANA workloads. The VMs will be powered by Intel Xeon Scalable processors, called Cascade Lake. This will make VM sizing and provisioning more cost-efficient for customers.
Intel Optane DC provides persistent memory with higher capacity. Enterprises will be able to scale up instances without unexpected costs. The offering will be available in alpha later this year.
New Pricing for Google Compute Engine
Many enterprises complain about the unexpected costs of a cloud infrastructure. This led to a growing market of cloud cost management vendors, like Cloudability. Many vendors have begun using pay-for-use models for certain cloud tools. Google Compute Engine becomes the latest to join this trend.
Their resource-based pricing model considers usage at a granular level. It evaluated how many resources you use over time, rather than on what machine types you use. This simplifies the pricing model to a GB of RAM is a GB of RAM no matter what. Also, they will calculate sustained use discounts based on region instead of within zones. Meaning enterprises can accrue sustained use discounts faster and automatically.
More Machine Learning Options
Access to Google’s analytics and data collection technology makes GCP standout. The company announced even more functionality for their data analytics and machine learning tools.
At Google Cloud Next, GCP announced BigQuery ML. This tool gives users access to the power of predictive analytics, even if they lack data science knowledge. GCP stores data for customers easily. Thus, BigQuery ML allows users to take advantage of their data when creating and deploying models.
Kubeflow v0.2 was also announced at the conference. The tool allows enterprises to deploy machine learning workflows on Kubernetes. It simplifies the management of Kubernetes in both scaling and portability. Since Kubernetes is open source, Kubeflow v0.2 helps enterprises deploy open-source systems for ML to diverse infrastructures. It eliminates the headaches of integration.