Hybrid Cloud IBM Cloud Migrate to the Cloud

IBM and VMware Transform Cloud Migration Through AI-Powered Smart Operations

This blog was co-authored by Josh Simons, Chief Technologist for HPC, VMware and Goldy Aloysious, Lead Architect for AIM, IBM


Artificial Intelligence and Machine Learning techniques have the potential to drive widespread transformations in IT. Learn how the AI-powered Cloud Migration (AIM) project is increasing the adoption of IBM Cloud for VMware Solutions by accelerating the migration process for enterprise customers who are looking to move to the cloud. 


The capabilities unlocked by AI and ML techniques have the potential to drive widespread transformations across society, including within IT. We seek to enable this transformation within IT in three main ways:

1.) Enabling joint customers to flexibly and efficiently run their AI/ML workloads on the IBM Cloud for VMware Solutions platform.

2.) Incorporating AI/ML into our products and services to deliver insights while increasing agility and scalability.

3.) Embedding AI/ML into our internal operations to improve efficiencies.

We are very excited to introduce our AI-powered Cloud Migration (AIM) project to simplify an extremely complex manual process. In this project, AI/ML is used to augment the power of migration practitioners, leading to a faster, more streamlined process.

This project has been sponsored by the VMware – IBM Joint Innovation Lab, and is intended to make cloud migration to IBM Cloud for VMware Solutions faster and more cost-effective, by accelerating and simplifying the migration process.

From end to end, the bulk of time during a cloud migration is normally spent in the discovery and affinity analysis phase. In this phase, identification of application topology and related affinity is mostly manual and therefore error-prone and time-consuming. Due to the dependency on clients to get the information needed for migration planning, the plan phase, discover phase and design phase take significant time.  

Often, management and maintenance of inventory data are done manually, which is suboptimal, particularly in cases of continued inventory changes. Therefore, the trust levels associated with inventory data are lowered while making migration decisions. The AIM solution uses AI capabilities to improve the productivity and effectiveness of an IBM Cloud Migration Factory (CMF) practitioner in the discover, design and migrate phases.

While in some cases the inventory of the client’s environment is gathered using discovery tools, there are many cases where the inventory has to be gathered manually by the CMF practitioner.  Even when discovery tools are used, their output is usually processed manually. The most significant tasks in this manual process involves looking for missing and inconsistent data belonging to the workload, identifying shared services in the environment, and identifying servers belonging to each of the many applications that are within the scope of migration by interviewing application owners in the client’s organization.

During the discovery phase, identifying bundles and deciding the sequence in which the applications should be moved, while also accounting for business constraints such as change window, business continuity and application dependencies, present additional challenges. A few other notable areas where AI can help cloud migration are; target environment optimization, advisory services and cloud migration opportunity miner.


AIM Solution Capabilities


Our AI-powered cloud migration (AIM) solution will optimize the discover, design and migrate phases of the cloud migration lifecycle by assisting Cloud Migration Factory practitioners to:

  1. Identify logical application groups, communities and shared services from discovery data, using AI community algorithms.
  2. Optimize the interview process by asking application owners in client organizations relevant, contextual questions that are generated based on the missing information.
  3. Optimize and automate the creation of workload bundles and wave plans.
  4. Optimize disposition recommendation engines using active learning, where the model training can be optimized with minimal labeled data sets. Workload classification uses machine learning techniques to automatically categorize workloads into migration bundles, such as rehost, rearchitect, retire and re-platform.

For the discover phase of the scenario mentioned above, the AIM solution assists the CMF practitioners in the following ways:

1. It allows them to identify the shared services in the discovery data using AI community algorithms. This will help save time by avoiding having to manually analyze the output of the discovery tool to identify such shared services. The shared services are identified and removed in the AIM solution from the discovery data using a two-pronged approach of maintaining an application signature for shared services and using algorithms to detect servers that have the highest number of input and output connections.

2. While shared services are important for building in the target environment, most application servers will have sessions or interactions with shared services servers, which are included in the output from the discovery tools. By filtering the shared services from the discovery data, the application server interactions are more visible to the CMF practitioner. The AIM solution, using AI community algorithms, logically groups servers (logical application boundaries) with frequent interactions, so that the CMF practitioner can use this information from various application communities for the initial meeting/discussions with the application owners in the client’s organization.







Affinity as discovered                              Shared service identified             Identification of servers as colored nodes


3. While other discovery tools may also logically group servers based on similar criteria, the AIM solution goes a few steps further. The AIM solution inspects the server nodes for missing data and analyzes application communities to generate contextual questions that can help the CMF practitioner during the interview process with the application owner. The CMF practitioner can also leverage this capability to standardize their interviewing experience with their clients.

4. In the migrate phase, the CMF practitioner would need to create migration bundles with migration events manually, which can be time-consuming. The AI planner component of the AIM solution uses existing bundling logic to optimize the bundling process, which can be fed back into tools like Transition Manager for the governance of the migration project. Future releases will have the capability to bring in additional dimensions such as the risk of workload migration, into consideration.

5. When applications need to be re-factored or re-architected when migrating to the cloud, the AIM solution has an active learning component that can optimize disposition recommendations by using a smaller sample of labeled data for training the models.

The above steps, using an AI-based solution, help provide CMF practitioners with the contextual information needed to enable their clients to make faster decisions on application topologies, thereby improving their ability to plan which applications to migrate.


IBM Services experts have already migrated and modernized over 100,000 workloads and applications to the cloud. This proven expertise, paired with our new AIM solution capability, will allow us to advise, move, build, and manage across the journey to cloud in an accelerated timeframe. Enabling AI-powered smart operations to consolidate the end-to-end process for cloud migration helps enterprises transform at the speed their businesses demand.


Let VMware and IBM be your partners in the cloud, and move to the cloud with confidence.