Why Anthos for the Enterprise
When Google introduced Anthos as a cloud offering in April 2019, they hoped to unlock value for their customers that was previously trapped in physical data centers. Customer flexibility, scale, reliability and mobility were among the key design principles that influenced Anthos. During the intervening years, Google has augmented the Anthos offering to encompass new products, centralized configuration and various deployment modalities.
One enterprise objective that Anthos designs for is to provide a single pane of glass and a homogeneous platform across clouds and on premise. From an operations perspective, this both reduces burden and increases responsiveness.
Central configuration management is also an integral part of the platform. Not only does this provide a seamless runtime approach between clouds and on premise, it goes further to ensure that security policies and access controls are consistent between environments.
Anthos provides several first class services, such as Google Kubernetes Engine (GKE), Cloud Run, AI/ML services such as Speech-to-Text, Anthos Service Mesh and BigQuery Omni. These, along with a growing list of other services, are all available regardless of where the compute is located, which can unlock the business value of your data.
What types of workloads does this unlock? Microservices; AI/ML; data analytics; serverless; advanced deployment, validation, testing, troubleshooting and delivery; and many more. Anthos also makes Marketplace offerings available wherever your compute is located, giving instant access to hundreds of predefined and supported offerings.
Why this matters to the Enterprise
In order to better understand the value available to an Enterprise in an offering like Anthos, it’s important to understand what many Enterprises look like. Companies that reach Enterprise scale often have decades of history, which means a wide variety of legacy systems. Some of these legacy systems came from mergers and acquisitions. Some systems may not have an internal champion or expert. Some may be kept around only to satisfy data retention. For greenfield projects within an Enterprise, there exists a tension between adopting new technologies and leveraging existing expertise and infrastructure. These factors all potentially complicate the path to modernization, which, not surprisingly, is what Anthos aims to deliver.
Focus on Business Outcomes
Path to cloud
Anthos offers a migration tool designed to minimize the effort and risk associated with application modernization and migration. At the same time, this approach doesn’t require teams to immediately change their delivery approach.
Avoid lock in to proprietary technologies
With Anthos, Google is showing a commitment to open source, both in principle and in practice. One benefit of open source is that it offers users flexibility about where and how they run their applications and serve their workloads. Anthos delivers that same flexibility to the enterprise, making it possible for them to choose the most convenient place to run their workloads, while maintaining a consistent application life-cycle. In practice, Anthos puts Kubernetes at the heart of its offering, which ensures portability and openness.
Satisfy data governance requirements while unlocking data for advanced processing
Some data needs to stay where it is. This may be due to data sensitivity, data sovereignty and legal requirements, to name a few. Anthos resolves the lack of data mobility by bringing the services to the data. Whether that’s BigQuery Omni or building custom kubeflow based workflows, data location no longer needs to determine what can be done with the data.
Maximize utilization across clouds and on premise
Low utilization is a tax on many enterprise environments. In some cases, the environments were oversized. In other cases the application utilization has diminished as business processes evolved over time. In yet other cases, development and test resources are consuming fixed resources, some of which are stale, that have accumulated over time. Whatever the reason, the containerization of these apps in Athos and Kubernetes can result in a significant increase in hardware utilization, and corresponding reduction in infrastructure cost.
Automation of policy and security across all environment
Enforcing a consistent security and governance posture is essential for most enterprises (and should be for all companies). Anthos designed a central configuration management function to allow administrators to establish security guardrails, enforce policies across Kubernetes clusters (e.g. RBAC, quotas, etc.) and manage cluster sprawl.
Many advanced traffic control and telemetry patterns are available when using a Service Mesh. These include canary deployments, traffic mirroring, request tracing and circuit breaking, to name a few. Each of these patterns increases the quality of delivery to the end user while mitigating possible adverse impact due to bugs and other unknowns that sneak through development and test. Anthos Service Mesh makes it possible to extend the mesh across environments in a fully managed solution.
Day 2 operations
Delivering software is so satisfying, but enterprises have to worry about more than just delivery. Software disruption can impact the company’s bottom line, which is why it’s essential to account for a strong operations practice. Anthos’ unique design reduces the labor required to maintain, patch and update VMs and physical servers. This is accomplished in part by way of modern CI/CD pipelines, a focus on image-based delivery and desired state configuration.
In addition to the apps you develop, you get access to hundreds of predefined and supported apps in the Marketplace. Many of these can deploy in a ready-to-use state with a few clicks. Obviously, since this is Anthos, these Marketplace offerings can be deployed across environments, wherever if makes sense for your organization to run them.
Deliver containers that run everywhere and stop worrying about hypervisors
With Kubernetes at the heart of Anthos, container based workflows are natively supported. Whether you build your own containers or bring a containerized app from a vendor, these can be deployed in a declarative way in seconds, without giving any thought to virtualization or underutilized resources.
Unlock additional capacity with bare metal deployment
Bare metal deployments give your enterprise access to 100% of the host system performance and capacity. Anthos on bare metal opens up new scenarios, like adding edge locations and supporting mission critical applications.
Anthos and the Modern Enterprise
Anthos means something different to each role in the enterprise. Below is a short list of some key roles, and why Anthos matters to each one.
Anthos provides a consistent experience across all environments, whether they are cloud or on premise. This consistency translates into a simpler operations playbook with more cross-environment skill overlap, which means fewer employees are required to maintain critical workloads.
Anthos’s commitment to open and interoperable technologies, like Kubenetes, means more flexibility in the enterprise. Whether sourcing from an ever growing list of ISV’s who support native deployment on Kubernetes or considering mergers and acquisition activity and wanting a straight forward integration between companies legacy systems.
Portability is a feather in the cap for the CIO who wants flexibility when it comes to deciding where workloads will run. Cost, reliability, data sensitivity and many other factors can result in a preferred choice among cloud providers or on premise. Anthos removes unnecessary work and mitigation from such decisions.
Security and Policy enforcement and consistency
Strong isolation for workloads
Single pane of glass, single point of configuration, single control plane for security across all environments
Advanced dev/deploy patterns
Low/no ops with high reliability, auto scale, etc.
Central configuration management
Burst to cloud
When to choose Anthos
Do you have data in data centers that’s captive (sensitive, PII, trade secret, partner agreement, data sovereignty, etc.)?
Do you have enough visibility into how your applications work, what data they send/receive, etc.?
Are your engineers able to deliver fast enough and work in collaboration with IT for a great experience? Are you able to use advanced deployment patterns, like canary and traffic mirroring?
Is your utilization high? Can you monitor and patch systems without disruption to apps?
Are your engineers able to use incorporate the latest AI/ML into apps that benefit your business? Are they able to bring these to production, at scale and with sufficient reliability?
What alternatives exist and how do they stack up
Amazon Outposts: increased data center footprint, hardware based solution
Azure Stack: Very little adoption and support
OpenStack: Heavy lift that requires staffing up and building a partner ecosystem
Strong Partner EcoSystem