Moving apps from the data center to the cloud
It’s widely understood that cloud-based environments can bring significant benefits to enterprises including improved cost savings and business agility. In fact, 82 percent of US businesses reportedly saved money by moving workloads to the cloud last year. In its 2014 State of the Cloud report, RightScale reveals that enterprises are adopting cloud computing in record numbers: Nearly all organizations (94 percent) surveyed are running applications or experimenting with Infrastructure as a Service (IaaS), while as many as 87 percent of organizations are using public cloud.
However, before businesses make the transition to a cloud-based environment, the challenges associated with the actual onboarding of enterprise application workloads are often overlooked. For businesses migrating to a cloud-based environment, ‘onboarding’ refers to the deployment of application workloads – or virtual machines and the applications that run on them – into a virtual private cloud or public cloud environment, without the need for major refactoring or rebuilding an enterprise application.
A significant challenge with cloud onboarding is that, for most enterprises, their applications are not designed for a cloud environment which is designed to scale on-demand and has failover capabilities built-in to the infrastructure. Rather, they are designed for the reliability of a dedicated environment.
Not surprisingly, the first applications enterprises typically choose to onboard are those already running in a virtualized environment. But not all virtualized applications are ‘cloud friendly.’ As such, enterprises must commit to early planning in order to prepare their workloads for cloud environments. This application onboarding process has 7 essential steps, which are outlined below.
Step 1: Define the workload
An application workload comprises all of the components required for the proper performance of an application. As hybrid emerges as the key cloud model, it’s increasingly uncommon for businesses to run single purpose, custom-built applications like Zynga and Netflix do. Rather, they use a combination of off-the-shelf applications like Oracle and SAP, which means their applications usually interact with one another across platforms. For instance, a logistics system which manages deliveries might be integrated with a CRM system. Also, not all of these elements will be migrated to a cloud environment – it’s very common for enterprises to maintain their Active Directory in-house, for instance.
For businesses considering migrating to cloud, they must undergo a thorough workload analysis early on. Preparing an application workload for forklifting into the cloud will need virtualization if the application is not already running in a virtualized environment. Ultimately, the nature and scale of the workload will dictate the number and type of virtual machines necessary for migration.
Step 2: Provision cloud resources
Servers, storage and network are required services for cloud environments, and service providers can help enterprises purchase these required amenities. To determine necessary resources, businesses should ask themselves: “What OS, databases and application servers are being used currently and how hard are they to migrate to the cloud?” “What are the CPU, memory, network and storage requirements and what will it cost to provide these in a cloud environment?” “What other software supports the workload?” and “What are the integration touch points with other workloads?”
Step 3: Establish a connectivity bridge
The connectivity bridge is the secure and transparent bi-directional connectivity between the data center and the cloud, which is essential for enabling cloud access. Most enterprises use Internet VPNs for connectivity to the cloud. However, Internet connectivity is not suitable for all applications. There are alternative, non-Internet connectivity options that enterprises can deploy, such as going through Ethernet (layer 2). Before choosing an alternative, enterprises will need to make sure their cloud providers can support them.
Step 4: Deploy the workload
Once the connectivity bridge is in place, businesses can setup their virtual machines and connect them to applications that remain in-house. Then, enterprises can transfer applications and any associated databases, software and services that they established from Step 2 as entrants for cloud migration.
Step 5: Ensure seamless two-way access
Critical to the success of any cloud migration is smooth integration between the cloud workload and the services not migrated, which stay on premise. With the hybrid model emerging as the dominant cloud environment for enterprises, IT will have more considerations, including connecting multiple cloud environments as well as managing enterprise-user devices that access applications in the hybrid environment. It’s therefore imperative that the two-way connectivity is fluid and secure and enables IT to monitor and manage the applications as well as the cloud infrastructure.
Step 6: Test and validate
To make sure applications will perform as required it’s important to conduct a trial period before launch. Businesses should ensure the application can recover from failure and test all third-party components. They should ask themselves: “Has everything been transferred correctly?” “Do network, storage, compute and database configurations remain intact?” “Can I see and manage the cloud environment properly?” and “Does my cloud backup process work?”
Step 7: Discontinue the old service
Once enterprises have identified and amended any performance glitches, they can then terminate the enterprises services and give all the users cloud access instead.
The new IT reality is hybrid
Today, more enterprises are moving to a hybrid model and beyond, into a multi-cloud architecture. This means IT is faced with entirely new challenges. Now, IT must navigate the complexities of integrating on premise workloads with those of a cloud service provider in remote, cloud environments. IT must also manage multiple clouds that comprise best-of-breed cloud services, including public IaaS providers and private PaaS, for instance. As complexities evolve, having a well-crafted plan, including workload analysis, making applications cloud-ready, and choosing the right cloud providers, will help businesses realize the potential of a hybrid and multi-cloud environment.
This article was originally published on Silicon Angle.