At Gartner IT Expo, Eyal Manor, VP Engineering at Google, walked us through a new Google offering for hybrid cloud.
Come learn about Google Cloud’s vision for hybrid cloud. This session will show you how Google Cloud’s Kubernetes & Istio-based Cloud Services Platform will bring the efficiency, speed, and scale of the cloud to you. We’ll help you see how Google Cloud is providing the future of cloud infrastructure; from the public cloud, to your datacenter with a consistent platform for modern application development. After this session you will understand Google Cloud’s vision for helping your organization build reliable, secure, and high performing cloud services for today, and the future.
Google previously announced their hybrid version that migrates many key services and technologies like GKE and Knative. You can also find their base offering page here.
Focus: how a modern hybrid cloud can help your organization move much faster.
Fact: 73% of organizations use cloud
Fact: only 10% of workloads have actually moved to public cloud
Fact: 80% of workloads that move to the cloud have some element move back. “If you can refactor an app it’s a great use case but if not, it’s a waste of resources.
IT Modernization means devops, built in security across the stack, portable, scalable, pay for use, and productive developers. Google believes public cloud brings most of the benefit but you have to take into account the journey.
Challenges: Getting from rigid classic apps into the modern cloud. Moving is harder than you think. It probably means refactoring your applications to use these services.
Options to move to public cloud:
- Rip and replace: really expensive
- Lift and shift: take what you have and just move it all into a vm on the cloud.
- It’s efficient
- You pay for use
- It’s more secure
- But you aren’t taking advantage of all the cloud services and scalablity
- Lift and Modernize
- Make some changes but not a complete rip and replace
- Improve and Move: e.g. use the hybrid cloud
- Modernize now
- Productive developers
- Portable to the public cloud when ready
- Scalable
- Available
- Assumes you are developing cloud native technologies
- Uses a range of cloud tools like Istio, Kubernetes, and Knative
- NY Times and Symphony have chosen this approach
The Cloud Service Platform
This allows you to shift from traditional infrastructure and IT centric. It brings things like auto-healing and lets you modernize your applications on day one
Metric: Improve time from code to production
CSP is optimized for agile development, hence the metric
GKE On Prem
This is the Google Kubernetes Engine. The supporting team created this and then open sourced it.
- turn key production grade
- easy upgrade path to latest Kubenetes releases
- Access to container services
GKE and GKE On Prem is consistent. You use the same tools to install, configure, and manage clusters in both.
Quote: IT’s the only systems that has the same api, tooling, and operational tooling anywhere.
You also get a single pane of glass to manage it. You see both Public cloud and On Prem.
Policy Management
This lets you provide role based policies and manage identity within the clusters. Y
Istio
It secures service traffic without additional code. It integrates with the rest of the open source tech stack. It’s becoming the industry standard since it was released.
ebay quote: Istio reduces our operational complexity by moving networking logice from our service to the infrastructure.
Key Bank: CSP provides us with the security we need with the portability we want and the productivity our developers crave.
HSBC: We needed a consistent platform to deploy and manage containers on premise and in the cloud. GKE was a natural for us
Who is using Kubernetes?
- Capital one
- Cleveland Clinic
- Bank of New Zealand
- Thomson Reuters
Takeaway
- You want to think about a consistent experience between public and on premise cloud
- You want centralized control that’s policy based across environments
- Agility with reliability brings productivity gains
- Flexibility to scale up or down as needs