Solving Cloud's Infrastructure Complexities With YotaScale Infrastructure Performance Management
When it comes to optimizing today’s complex cloud environments, it almost seems that you need a Ph.D. in cloud computing just to understand all of the moving parts. There are countless variables on the infrastructure side to consider that are relevant to specific workloads and the KPIs that they need to meet. Consider this: there are about a hundred thousand combinations when spinning up a web server in AWS. And that’s just a web server. For today’s typical enterprise organizations, there are likely millions of infrastructure combinations available at any given time.
There are simply too many options, too much data, too many moving targets and too many decisions to make. How do you find the best combination of infrastructure for all of your organization’s unique needs? And how can you accomplish this in real time?
This is the problem that Asim Razzaq, co-founder and CEO of YotaScale, ran into when he was PayPal’s senior director of engineering responsible for all of PayPal’s core infrastructure. He knew that other enterprise companies would also be struggling with figuring out all of the different cloud infrastructure variables. He decided that there needed to be a new breed of software that can handle this type of enterprise-centric management work.
AI Powered Infrastructure Performance Management
In 2015, Razzaq co-founded YotaScale, which offers AI powered tools for optimizing cloud infrastructure for things like performance, efficiency, availability, security and costs. It’s a software-as-a-service offering currently running in AWS.
Up until now, there was no real solution for infrastructure performance management for the cloud that could optimize your cloud computing environment based on all of the available data points.
“There are a lot of tools out there that do the porting and the business intelligence,” says Razzaq, “but our customers actually end up calling them “data spray” tools. They’re spewing more data at them that now they have to sort,” he explains, adding that they’re getting additional data that they’re unable to keep up with.
Razzaq says that his ultimate vision for YotaScale is self-managing infrastructure. The CEO says that currently, YotaScale focuses on three key things: anomaly detection, continuous optimization, and contextual analysis. What’s unique is that YotaScale has insight through the AWS API, so it’s able to pull the real costs of instances, not estimates.
Anomaly detection, one of YotaScale’s flagship modules, looks at different variables like cost, utilization, performance and other parts of an organization’s infrastructure in order to bring knowledge to the surface, Razzaq explains. By examining historical data the tool is able to tell what is abnormal behavior in a certain part of the infrastructure and what the root cause of the problem is.
The tool is intelligent in the sense that it’s able to look at several things at once to determine if the instances are right sized and performing adequately. It might suggest that you size an instance down to lower cost, or size up to improve performance.
“We also do predictions for the future by leveraging machine learning and core relationship algorithms,” says Razzaq, adding that a lot of their customers find anomaly detection extremely useful because it can take someone days just to sift through all of this information.
YotaScale’s unique advantage is that if you already have historical data in AWS, the anomaly detection tool can use that as a baseline. Once an anomaly is identified, the system makes recommendations on remediation and what actions need to take place. The vision, Razzaq mentioned, is that the remediation will be automated, but that’s a feature that’s still in the works.
The continuous optimization module focuses on identifying the best combination of infrastructure to deliver on the right KPIs for a particular workload.
“Continuous optimization is about bringing best practices in the context of the enterprise,” says Razzaq. “Best practices are changing; the infrastructure combinations are changing; my workload is changing. Our goal is to bring all of that to you.”
Contextual analysis is about enabling visibility through a cloud optimization “score” that compares your workloads to similar types of workloads at other companies.
“It’s very interesting to me that a ton of companies out there don’t even have a baseline for cloud efficiency,” says Razzaq. “We have a number of algorithms that go and surface that efficiency to say we believe you are at a 56% compared to other companies that we’ve seen or compared to all the best practices that are out there. And here are the things that you can do to get better.”
YotaScale is basing this knowledge on data from different companies with similar workloads comparing it to companies that are able to accomplish the same type of workloads more efficiently, Razzaq explained.
“Most of these best practices right now are tribal knowledge,” says Razzaq, “They’re on blogs or whitepapers and they’re static in nature. You have to read through them to implement them.”
The contextual analysis tool aims to bring these best practices to organizations so they can be easily implemented as things change.
Getting Started With YotaScale
YotaScale doesn’t require an agent. Instead, it gains insight into workloads through what Razzaq calls boundary detection, leveraging an ecosystem made up of other providers and tools like CloudWatch, New Relic and a number of configuration management tools.
“We have some unique techniques that do boundary detection without having to go too deep in the system,” says Razzaq.
Because YotaScale is a SaaS platform, getting started with the service is fairly simple. There’s no software to install, no hardware to manage. It’s a service level account that just needs the right permissions to get started. And each company has full control of what permissions are granted based on their comfort level, Razzaq explains.
Now for all of you Star Wars fans out there, you’ll be interested to know that the name of the company was actually inspired by Yoda. Razzaq tells me that the unit yottabyte (the symbol for it is YB) was based on the popular Star Wars character and that’s where the “yota” in YotaScale comes from.
To learn more about YotaScale and get access to a free 14 day trial, visit yotascale.com.
The ActualTech Take
It’s no secret that complexity in infrastructure and applications today is rampant. Nor is it lost on anyone that all the data we collect does have meaning, but it’s elusive. Harnessing the power of AI to extract the signal from the noise is a shortcut to understanding where real problems lie and how to optimize environments. When it comes to creating meaning from data, context is everything. Therefore, a tool that captures and correlates important relationships between metrics and also compares results to other deployments helps bring some clarity to IT organizations.