“Cloud First” has been the mantra for years, and for many organisations, it has been an evolution of abstraction. It started as “Virtual First”, with the abstraction of logical servers away from physical servers. “Cloud First” then evolved into the abstraction of servers from physical data centres. Most recently, “Cloud First” has become the abstraction of applications from middleware and databases through Platform-as-a-Service (PaaS).
Through the iterations of “Cloud First”, we’ve also seen the definition of cloud evolve, and many vendors are now blurring the lines between where public and private cloud capabilities exist. It’s now possible for clients to have capabilities in their own data centres that were traditionally only available in a public cloud. Meanwhile, the single-tenancy of a private cloud is now available within a public cloud. Further blurring the lines are products quickly gaining market share across all industries that allow for workloads to be developed once and moved from on-premise private clouds to a public cloud or from one public cloud to another. With more choices than ever, it has become an increasingly complex universe of products and services that are difficult to navigate.
Bringing the public cloud on-premises
One of the first products to be widely released has been Microsoft’s, Azure Stack. Available to customers as a black box, Azure Stack is a complete package of rack(s), servers, storage, network and SAN switches and software. Users don’t have to manage underlying hardware, software, management or monitoring. It’s just a simple-to-use web interface for admins and developers. Azure Stack’s ideal use cases are for edge (local data processing later aggregated to central locations), disconnected (poor or no connectivity locations often found in manufacturing), or regulatory scenarios. Azure Stack provides the ability to deploy servers, containers, Database-as-a-Service (DBaaS) and PaaS web services in a consistent manner, so developers can write the code once and deploy it across multiple locations. Azure’s public cloud capabilities in a hands-free, on-premise solution also allow applications to be easily moved or extended from on-premise to public cloud with ease.
At Cloud Next 2018, Google announced its GKE On-Prem offering. In early access, GKE On-Prem brings a validated, tested and managed version of Google’s Kubernetes Engine on-premise. Offered as a software solution, it allows users to have GKE On-Prem deployed on top of existing hardware and private cloud environments. While GKE On-Prem offers a more limited set of public cloud features, it gives customers a consistent Kubernetes experience to run containers in sync with Google Cloud Platform (GCP) versions, use the same Google Cloud Console for centralised multi-cluster management, and provide easy portability of applications between on-premise and the public cloud. In addition to its integration with the Google Cloud Console, GKE On-Prem also integrates with GCP’s Identity and Access Management (IAM), Stackdriver monitoring and logging, Google Cloud Build and other GCP services. GKE On-Prem offers users the ability to write and manage containerised applications in a simplified and singular method to combine private and public cloud.
At Amazon’s massive re:Invent 2018 conference, they announced their upcoming AWS Outposts offering (release date unannounced but expected in the second half of 2019). Similar to Microsoft’s black box approach, AWS Outposts will be a fully managed hardware and software bundle that will provide an interface similar to the AWS public cloud. At launch, AWS Outposts will only support IaaS server workloads through Elastic Compute Cloud (EC2) virtual machines and Elastic Block Store (EBS) file systems.
Taking private cloud public
VMware still holds the top spot for on-premises hypervisors (91% according to Gartner’s Market Share Analysis: Virtualisation Infrastructure Software, Worldwide, 2017), and most organisations have built robust operations groups around their VMware capabilities. Amazon and VMware partnered to provide VMware Cloud on AWS as a way for enterprises to move their existing workloads to single-tenant servers. Existing teams can manage and deploy workloads using familiar processes and tools, reducing the need for additional training.
Workloads can migrate between on-premise VMware instances and VMware Cloud on AWS for seamless portability. This common architecture is also allowing enterprises to quickly and confidently deploy, configure, and test DR. New innovations are enabled through VMware Cloud on AWS’ direct connectivity into the AWS public cloud where cloud-native services can be integrated and called by the legacy applications still hosted on VMware.
Hybrid capabilities across platforms
Pivotal Cloud Foundry (PCF) is built upon the open source Cloud Foundry, and provides additional, enterprise-level services and fully managed support. PCF is based on virtual machines and containers, which allow organisations to create their own PaaS services, or deploy PaaS services from their Pivotal Services Marketplace. While PCF does add a layer of management, it also provides the enterprise full control over their PaaS environments and data. The use of containers, PaaS services, a shared marketplace, and integration into popular DevOps tools makes it an attractive option for many organisations looking to develop cloud-native apps without the tie to a specific cloud service provider. Organisations can deploy PCF to on-premise or public cloud environments and gain the ability to migrate workloads between cloud models and cloud service providers seamlessly.
Kubernetes(K8s) is a container orchestration engine originally developed by Google and then donated to the Cloud Native Computing Foundation as an open-source package. Kubernetes allows users to deploy software as containers which are then managed, load balanced and auto-scaled. Although originally developed by Google, Kubernetes has become the standard for container orchestration on-premise and in the public cloud. Microsoft and similar companies have recently announced the discontinuance of their own container platform, such as Azure Container Service (ACS), in favor of a Kubernetes based solution. Many popular container and PaaS platforms are also built upon Kubernetes; including Azure Kubernetes Service (AKS), Google Kubernetes Engine (GKE), Amazon Elastic Container Service for Kubernetes (EKS), Pivotal Kubernetes Service (PKS), Platform 9, RedHat OpenShift and OpenStack. As a platform-agnostic container orchestration engine, organisations are able to develop containerised applications on-premise, and then move them to the public cloud while still retaining a common development and management process.
At NTT DATA Services, we believe organisations are yet to realise the full potential of cloud computing. Cloud today is usually viewed by enterprises as part of a larger strategy around digital, transformation, costs savings and/or entering new markets.
Our Cloud Advisory Team helps clients weave cloud into the core of their businesses and deliver everything related to a cloud transformation – the first step is a cloud first approach.
Post Date: 12/03/2019