Decades of collective migration experience embedded into our purpose-built junkshon workload mobility analytics platform
Achieve results in days not months
junkshon's seven step process
Once loaded, the data is analysed for patterns and gaps by comparing it to junkshon's library of exemplar data. The learning algorithm is used to fill gaps with weighted certainty to inform later steps in the process. At this stage junkshon can also filter out applications determined to be out of scope or left until later waves of migration.
Migrating tightly coupled application components and related instances in the same wave is key to reducing performance issues, latency, downtime required and impact on dev and test teams. junkshon's grouping algorithm generates the optimal groups of applications and configuration items into each wave.
Application Dependencies and Surveys
Dependency mapping is the second step in the process. A lot can be determined from a little data. The more junkshon can understand about the application dependencies (libraries, services, startup routines etc.) and interfaces (external data, databases, ecosystem etc.) the more advanced and automated migration execution treatments can be used.
Not all migrations to cloud should be lift and shift into IaaS. Finding the highest ratio of benefits to change effort is achieved by matching a hierarchical list of migration types to your applications. Preferred migration types are: App to SaaS, Containerisation and web, middleware and database to PaaS. New cloud services keep being launched which is why we keep updating our migration types and tuning junkshon's algorithm.
When junkshon has understood each application and infrastructure instance, a series of weightings are given to each one. The weightings include complexity, risk, uncertainty, BU criticality, change agent needed, cost and reward factors and rate of change. The weightings create a benchmark score for each instance. In the later steps the benchmarks are added together in each group to create a seeding for the scheduling algorithm.
Most infrastructure is under-utilised. One of the greatest contributors to cost savings is resizing when you migrate to cloud with its elastic scaling capabilities. junkshon contains a right-sizing function to plan the capacity needed based on existing usage data. AWS, GCP and Azure have a plethora of compute and storage offerings of different price points and performance attributes. junkshon maintains a realtime link to multiple cloud service catalogues to feed the right-sizing and pricing calculation.
All of the previous steps come together in the sequencing function to produce the migration schedule. junkshon's migration algorithm recursively generates a weighted sequence of application groups by change window. A host of business, operational and technical variables can then be used to refine the plan as the migration programme is executed. Variables such as business criticality of application, release cadence or change freezes can be used to tune the order of the plan. Technical factors such as location, platform, DC facility exit plans or end of life upgrade projects can be used to reprioritise the schedule. The schedule can also be refined to ensure a percentage of apps move to cloud or a certain date, a saving figure or acceptable risk profile is achieved.