Stewart Hunwick is the field chief technology officer (CTO), for cloud platforms, in EMEA, at Dell. We discuss artificial intelligence (AI), and its use in storage in this podcast.
We talk to Hunwick specifically about artificial intelligence (AI) for IT operations as a tool that can improve storage management, sustainability, and data security.
Hunwick likens
AIOps can be used for everything from storage management to a fitness tracker in the datacenter. It can reduce the carbon footprint of a fleet and monitor energy consumption at micro and macro levels.
It can also enable proactive and reactive responses to security concerns and allow administrators to ask natural language questions about devices under their control and even generate code.
What is AIOps (Analyze, Organize and Operate)?
AIOps stands for artificial intelligence operations. It’s a way to use this amazing burgeoning AI field to streamline and enhance your operations. It’s like a fitness tracker for your datacentre.
This is a great comparison, because a fitness tracker gathers raw data about you, crunches the numbers, and then gives you the insights.
AIOps is a concept that focuses on understanding the metrics of a datacenter and using those metrics to predict what will happen in the future.
Identify anomalous behaviorIt can also be used to reduce risks, increase sustainability, improve performance, and simplify operations.
What can AIOps do to help with storage sustainability. Sustainability is a concern for many organizations, and we can approach it in a variety of ways.
When we look at this specifically from the storage point of view, we see that organisations are trying to reduce their carbon footprint as well as the power consumption of their storage platforms.
AIOps is all about understanding the current power consumption, not only at a micro-level, such as per system, but being able to view it at a global level. So I can see my carbon footprint, my energy consumption, across my entire fleet of datacentres.
Then, using this data, I can start building in these trends, those projections, so that I can say “OK, here is what my current power consumption is.” I see an increase of 5% year over year. So, I can tell that something has changed.”
I may also be able to say “Well, this system is showing a very unusual spike.” This system is acting a bit strangely, which could indicate that a certain workload is consuming more performance than should be.
We can go further and project what the power consumption or carbon footprint requirements would be in future. Again, we can use all of that data to model what the future might look like.
It’s important to not only understand what you are currently using and consuming, but to also plan for the future. You can then make decisions about where you want to place new workloads and which systems will be deployed, and you’ll have a solid idea of how the future might look from a power consumption and usage point of view.
How can AIOps assist with data security? Data security is an extremely broad topic. We’ve had great success with many organisations in different verticals. In particular, we’ve had success with the healthcare vertical where we can simplify administration through this AIOps approach.
We can think of security in two ways: reactive security and proactive security.
AIOps can help in a proactive manner. It can tell if the data access patterns are normal or abnormal because it is able to see them from a data perspective. It can check things like the data reduction ratio.
If the data reduction ratio suddenly drops, it’s a sign that the data being written is pre-encrypted, or pre-compressed. If the data is pre-encrypted then it may indicate that a ransomware infection is in progress.
By using this metric in conjunction with the rate of growth of data, where the attack is coming from, we are able to pinpoint the source and say “Hey, it’s happening.” This is the reactive side. We know the source and we are confident.
A real-time attack is taking place. Let’s find out where it is coming from.
If you can imagine, as you leave for work each morning, having an app on your phone to quickly check that you haven’t locked your windows, or that you haven’t locked the front doors, or that you’ve left your keys in the backdoor, would be a proactive way of checking. Telnet, for example, can be enabled on a computer. This is because it has no encryption, so this could be a vulnerability.
AIOps can scan the configuration of your systems to identify weaknesses. We can also identify how the systems are configured, and spot attack vectors even before they are used. Checking if the door is locked before anyone attempts to break in.
What can AIOps do to help with storage management.
The topic of storage management is a hot one. Many organisations want to talk to us about this topic. It comes down to the same example that I gave earlier in the healthcare industry of driving down costs and simplifying administration.
The most recent growth is in taking AIOps a step further by leveraging generative AI, and building in an agent that can get to the heart of a question quickly. You might ask, “How does that feature work?”
or “What is the benefit of this particular capability?”
As an example, a metro replica. You may then ask, “How do you set it up?” or, “Give a brief summary of this particular capability so that I can send it to my leadership team.”
[You might say]”I need to create an Ansible Playbook to deploy in our environment.” It will then come back with “All right, Here’s a Sample Playbook you might want”
The most recent innovation is the ability to ask questions specific to your environment. Instead of [just being] being relevant to documentation, the ability to ask, “What’s going on with this particular system?” is a much more useful feature. Is there any performance impact on this particular system due to replication? How can I fix the system?
It can do all that stuff because it understands all the data and can see the context.
