As a vCloud Air Network service provider, building and offering hybrid cloud services to customers based on the SDDC is only half of the battle. Making sure that they are able to consume that service with fluidity becomes a critical area of focus. The less friction this Cloud Migration process has, the faster the customer time to value and service provider time to revenue become.
To address this rather broad subject, VMware is publishing a new document, “Migration Strategies for Hybrid Cloud,” in the VMware vCloud Architecture Toolkit™ for Service Providers (vCAT-SP). This blog introduces high-level concepts from that document. These concepts are meant to help both service providers and customers alike understand the opportunities and challenges when undergoing migration to a hybrid cloud scenario. Because this topic covers a vast area of information, the document only covers a few of the use cases available. Many of the more advanced use cases are accomplished through VMware Technology Partners, so stay tuned to the vCAT blog for additional information on how these solutions can be leveraged for migration to the hybrid cloud.
Figure 1. High Level Tool Categories for Migration
Looking at the figure above, we can see there are four main categories of tools available to accomplish different phases of the migration workflow. While there is a possibility that each category will be provided discretely in a single tool, it is often the case that single tools function in more than one category. It is also quite likely for most migration use cases that operators must coordinate activities between the tools in a workflow. Some, or in a best case scenario, all of these capabilities are integrated to help parties carry out a significant number of steps depending on the variables required for each migration instance. Leveraging the SDDC and its APIs provides the opportunity to automate as many of these steps as possible, and many of the available tools will facilitate some level of this type of automation.
More often than not, however, the governance of the migration projects, or perhaps even programs, should be addressed with a Migration Center of Excellence. In this Migration COE, typically hosted by the service provider, will be one or more instances of this tool chain constructed to allow customers and potentially other partners to come together and understand all of the variations that may drive migrations. Too often there is a rush to the workload migration tools themselves to relocate applications to the cloud without having considered the pitfalls, risks, and even upside potential offered by looking at the problem holistically. Specifically, we want customers to leverage the SDDC along with other VMware and Technology Partner solutions to introspect current application architectures as well as plan, and perhaps automate, the target tenant consumption of the service provider. The Migration COE allows us to visualize the “best fit” combination of tools and processes to plan the customer migration experience. The more information that can be applied to the process, the better.
By gleaning all of the potential information about source infrastructure and applications, we can create a repository of knowledge to plan migrations. The more virtualization and cloud-oriented solutions that are installed on the customer premises, such as VMware NSX® or VMware vRealize®, the more “migration ready” applications under the management of that infrastructure become. This is due to the ubiquity of the target hybrid cloud architectures, both built on the VMware SDDC. The primary function of the discovery and assessment tools is to ascertain the dependencies of the applications at a functional technology level. Examples of this might be DNS, PKI, or other authentication/authorization services, such as LDAP, that need to be made available to the application in its new post-migration home. Determining these dependencies will go far in planning for the serialization and parallelization of follow-on tasks related to migration and help to feed the three downstream task types—job scheduling, workload migration, and application verification. A great example of this customer-centric approach to discovery and assessment leverages VMware NSX and VMware vRealize Log Insight™. Once configured, the solution provides visualization of network activities through the Log Insight NSX for vSphere Content Pack v3, including application component interaction through networks and ports as described in this video.
Another important topic discussed in the migration document is workload mobility. There are a number of ways to provide hybrid cloud network connectivity (some are described in the blog Streamlining VMware vCloud Air Network Customer Onboarding with VMware NSX Edge Services), and many ways in which customers understand the concept of workload mobility. Because of the SDDC abstraction, many concepts discussed in the vCAT-SP use the terms “underlay” and “overlay”. While there is an obvious requirement for Layer 3 network connectivity to each site, the architecture will depend on the VMware software capabilities available at each site. Customers may choose VMware vSphere® metro clusters, a disaster avoidance scenario using VMware vSphere Replication™, or disaster recovery with VMware Site Recovery Manager™.
The Migration COE may include recommended methods based on any or all of these capabilities to help understand which may be appropriate in what situations. The hybrid network types in the previous paragraph provide workload mobility in the SDDC portion of the underlay that require VMkernel ports and operations. Migration solutions discussed in the vCAT-SP migration document, however, focus on the overlay consumption of hybrid cloud networks provided by VMware NSX in the creation of target environment capabilities to facilitate acceptable application characteristics in the new hybrid cloud location. For example, the creation of VMware NSX Distributed Firewall policies for application-centric micro-segmentation as described in the vCAT-SP blog, Micro-Segmentation with NSX for vCloud Air Network Service Providers. Because the overall costs of labor in a migration can exceed 50%, as described by pro forma cost model in this Forrester brief and detailed in this blog, migration becomes the lynchpin for the entire process of acquiring and recognizing new customers consuming the services offered. Choosing the right combination of tools and labor then is in the critical path to making sure migrations function in an optimal fashion.
Another critical facet that that might be outside of the Migration COE is capacity planning. The different methods used in workload mobility require specific underlay network capabilities to achieve their goals, mainly bandwidth/throughput and latency. More information on underlay networking for hybrid cloud can be found in the vCAT-SP document, Architecting a Hybrid Mobility Strategy with VMware Cloud Air Network. It is important to understand that the entire phenomenon of workload mobility, including migration, is a numbers game and not just of network performance. The customers will demand an understanding of how the application will be managed for performance and maintenance in the new environment, perhaps through SLA’s, which will be used to forecast the service provider’s TCO of the hosting environment. Provider compute/storage/network infrastructure must be provisioned in time to accommodate new tenant migration activities including potential shared transfer storage along with ongoing performance requirements. Some of the main drivers for the application cutover itself can be related to Recovery Point Objectives/Recovery Time Objectives, perhaps requiring the introduction of a hardware storage replication scheme into the mix. Consider also operational lead times for deploying and making these items ready for consumption and the potential ROI from automating as many tasks as possible.
Finally, one of the key reasons a service provider would drive their customers to collect the fullest amount of data possible is to leverage it to predict which customer workloads come with a “stickiness” to new services offered by the service provider and their partners. The ability to digest and manage all of this data in an effective, holistic way provides agility, creating a migration “funnel” of activities, fully leveraging but not exceeding capacities. This is achieved while also sustaining transparency to stakeholders, which is very powerful when a new journey is undertaken. Because vCloud Air Network offerings are built on the VMware SDDC you can be confident that it will offer the greatest compatibility and ease of both migration and mapping new operational procedures based on best practices in the vCAT-SP.