Unravel Data Adds AI to Prevent Cloud-Migration Cost Hangovers
For all the benefits of moving to the cloud, cost is increasingly considered part of the untamed frontier. It’s even been called the “cloud hangover” among companies shocked by their bills.
“Making the internal shift from a CapEx datacenter operating model, to an OpEx variable-spend model has taken many IT, finance and business teams by surprise,” wrote J.R. Storment and Michael Fuller of the FinOps Foundation.
Determining cost is just part of the complexity of moving Big Data applications such as those using Spark or Hadoop to the cloud, according to Unravel Data CEO Kunal Agarwal.
His company has found that its full-stack data approach, originally targeted to application performance management, also can be used to help customers more intelligently embark on cloud migration.
Unravel uses machine learning and artificial intelligence on top of log data to monitor application performance, detect anomalous behavior and offer recommendations to remedy the situation.
It announced its cloud-migration capabilities back in April. Now it’s touting a cloud migration assessment combining its technology and expertise, including new models specific to moving to Azure, AWS or Google Cloud.
Companies moving to the cloud typically face some common problems, Agarwal said. They typically don’t understand the cost and effort it will take, and underestimate the time it will take, with either the project going slowly or failing outright as they move from one environment to the other.
Some customers have already had projects fail, but want to try again, he said.
“But why would you want to start another dumpster fire by doing the same thing over again?” he asked.
AI-Assisted Cloud Migration Planning
The service starts with an AI-based discovery phase of the existing environment, a process typically done manually at most organizations.
“We gather this full-stack data – we used to do it for performance monitoring and optimization. Now using it to help companies understand which apps they should move to the cloud. If you move one app or an entire cluster to the cloud, how much will it cost you? How do you do a comparison of performance, compatibility, features across the various cloud options out there? Then really helping these customers understand the dependencies. What kind of compute and resources will be needed?” he asked.
The new models are designed to help customers understand which cloud provider is best suited for a specific workload, then which instance types and VM types with that cloud provider are the best fit.
The assessment can be used with leadership to make the case for the cloud: How much will we save? How much agility will improve for us? It enables cloud capacity planning and chargeback reporting, as well as other critical insights.
The service identifies workloads suitable for the cloud, determines the optimal cloud topology based on business strategy, and calculates the anticipated hourly costs. It also provides recommendations to improve application performance.
It can be used as project management tool to track your progress. You can use Unravel to determine performance and cost baselines.
Then it allows you to create a project plan: which pieces to move first? Afterward, you can use it to assess whether you achieved the quality of service, performance and cost you had hoped for. If not, you can use Unravel’s optimization features to get the project back on track.
Unravel won’t completely replace other tools and people in this process, Agarwal said. For read/write applications, refactoring applications, moving the data lake and similar tasks, you still will need other tools and people.
“But you’ll know what you need to do and not be left blinded by any step of the process,” he said.
The company recently announced partnerships with Informatica, Attunix, Dell and RCG Global Services to help customers with their cloud migration challenges. M12, formerly Microsoft Ventures, is among the company’s investors.
Feature Image from Michal Jarmoluk and by Clker-Free-Vector-Images, from Pixabay.
