Plan large-scale IT migrations: Cloudifications, Relocations, M&As, Splits, Replatforming, etc.


modelizeIT Value Add for Cloud Migrations Planning
Discover applications, their architecture, and dependencies
unknown unknowns
Enterprise IT environments are complex. Some information is well documented. Some information is known to IT staff and can be collected via interviews. However, some knowledge is always lost and nobody even knows what is not known. With our automated business application topologies discovery capabilities we are going after these so-called unknown unknowns.

After collecting the data modelizeIT platform auto-identifies groups of servers that logically work together and likely form business applications. It also generates architectural diagrams for each such group allowing migration architects to accept proposed business applications mapping with a couple of clicks or refine the results.
Deep information about middleware systems, deployed applications, databases, clusters, middleware cells, and auto-recognition of custom/rare applications makes it possible to identify the applications even if no traces of their documentation or IT personnel knowledge is available.

modelizeIT platform allows IT personnel and migration architects augment auto-collected information with minimal amount of manual labor.
Learn More
Identify unused servers and software and reuse/retire instead of migration
Non-trivial amount of servers and enterprise software in datacenters are not used and drain IT budgets with no positive business value. modelizeIT platform analyzes application architectures and usage patterns to auto-identify such unused servers so that they can be retired or reused instead of a costly and sense-less migration. Learn More
Estimate capacity and costs
Enterprise-class capacity and cost estimation is a multi-dimensional problem:
  • We discover enterprise software-licensing-related information. Enterprise software licensing metrics depend on the details of hypervisor configurations. Therefore, the same software installation may cost dramatically differently depending on the specific hypervisor, cloud provider, enabled options, etc. Moreover, common migration tools cannot migrate certain OS-dependent software installations and it is important to identify them in advance.
  • We discover complex storage topologies. For example, some storage volumes may be used by databases directly and be visible to many servers (but not be mounted from any specific server). Without taking this information into account storage-related capacity and pricing estimates may be significantly off.
  • We estimate CPU capacity taking differences between CPU architectures and virtualization technologies into account (e.g., one vCPU on source LPAR with shared uncapped CPU pooling is not equal to one vCPU in a Linux-based cloud).
Architect target solutions
Many migrations require changes in the IT architecture: clustering, backup, load-balancing, security, and many other mechanisms change dramatically during cloud migrations. Similarly, a client-managed database may become a cloud-provider-managed after the migration. A database instance listening on the same port now may be split between two different servers so that multiple applications each have their own dedicated database cloud image after the migration. In any case, it is important to understand the current architecture of each business application before making the target architectural planning. However, the current enterprise architecture is always complex and non-trivial. We model enterprise IT architecture automatically and capture it the way it is, with no trivialization assumptions (e.g., that each server belongs to some specific business application or that a cluster contains servers). Our auto-generated diagrams serve as the basis for the enterprise solution architecture planning and cut the architecture planning efforts and reduce the risks of mistakes.
Non-Trivial Clusters
Example real-world non-trivial clustering architecture where database instances are clustered differently among four different servers.
Plan migration waves
Large-scale migrations cannot happen at once: the work must be planned and divided into waves. Servers and applications depend on each other for proper operation. Just imagine how much effort is necessary to manually collect correct information about tens of thousands or millions of the inter-dependencies and analyze them all to make the division decisions. Our automated waves planning algorithm analyzes server and application inter-dependencies and replaces manual labor that is lengthy and error-prone to:
  • Accelerate migrations planning
  • Reduce the risks of unplanned downtimes (we generate wave and server dependency diagrams so that migration teams can see who gets affected)
  • Accelerate migrations and reduce their costs (optimal division into waves reduces the number of costly testing iterations due to the application inter-dependencies).
1000s of Job Scheduler jobs split into waves for migration
Example set of servers divided into waves based on the analysis of thousands of jobs orchestrated by job scheduling software running on these servers.