Table of contents
Deploying software to support the operations of an enterprise is an increasingly complex task that is often referred to as devops. When enterprise teams began using artificial intelligence (AI) algorithms to more efficiently and collaboratively manage these operations, end users coined the term AIops for these tasks.
Large software installations may benefit from artificial intelligence (AI) by monitoring software performance and flagging any anomalies or instances of poor performance. The software can review logs and track key metrics, like response time, to assess the speed and effectiveness of the code. If the values deviate, the AI can suggest solutions and even implement some of them.
The process involves several stages:
- Detection or observability: The software absorbs as many metrics and event logs as possible. The focus is generally on poor performance that can affect users directly, like a 404 error or an especially long database query run time. Some systems, though, may watch for other issues like a failed sensor or an overheated device.
- Predictive analytics: After collecting data for some time, AIops software can begin to identify precursors that can often signal an upcoming failure. The AI algorithms are optimized to look for correlations between values, especially those that are anomalies that may indicate upcoming problems.
- Proactive mitigation: Some AIops algorithms can be tuned to respond immediately to potential problems when the solution is straightforward. For example, a crashing service may be rebooted or reinitialized with more RAM. When these solutions work, they can eliminate much of the problem and save end users from encountering failures.
AIops is becoming more complex as teams deploy algorithms to a wide range of organizations. One of the most valuable opportunities arise when organizations begin to utilize other AI algorithms in daily operations. In these instances, AIops can help with deploying AI. This way, there may be synergy between the software layers.
Sometimes AIops teams use other subterms for their work. For example, MLops deals with using and deploying machine learning algorithms. DataOps can refer to the general problem of collecting data or the more specific problem of organizing data that is used to train and refresh an artificial intelligence model.
Also read: MLops vs. devops: What makes it different?
What are the options for AIops to support the deployment of artificial intelligence?
When AI scientists began to research AI algorithms, they worked with experimental machines in their labs. Now that AI is becoming regularly deployed in production environments, some are beginning to specialize in maintaining and maintaining software.
The difficulties in supplying AI algorithms are the same as maintaining regular software. There should be sufficient computational power to meet all requests, even those that come together in a time of peak demand. There should be a mechanism for testing changes and eventually replacing the software on the front-line machine with the latest version.
Although most of the work is similar to standard devops, there are still issues that pertain to AI and machine learning (ML). Some examples include:
- The model is like another piece of software with its own version number and history. The AIops team will juggle models, often independently of the software itself.
- Training the model is often a time-consuming process that often requires an elaborate build process of its own.
- There are now different chips that are optimized in different ways for creating the model and running the model in production. AIops teams must plan the best available hardware for each task independently.
- The build process may involve much more experimentation than typical software development. Its not uncommon for AIops teams to try different arrangements for neural networks and then evaluate how they perform.
- AIops teams may also have a third job of tracking the datasets that are used for training and evaluation. These datasets may also evolve with their own version numbers and history.
- Some applications deliberately feed data back into the training set over time, so the set grows and the results improve. AIops teams must also maintain the evolution of the training data over time.
- Some AI applications require screening results for potential bias. AIops teams can watch the working results for potential problems.
All of these questions and strategies apply in some way to the subsets with the names DataOps, MLops, ModelOps, and PlatformOps, because they focus on some of the specific tasks.
From Star Wars to streaming wars, here's how AIops is causing the intergalactic streaming battle.
Is AIops a mix of both?
Some businesses focus on using artificial intelligence to enhance the performance of their servers and databases. They refer to using artificial intelligence algorithms to monitor anomalies and, perhaps, anticipate outages or failures before they happen. The algorithms are capable at developing forecasts and alerts when the stack begins to perform differently.
AI algorithms are particularly useful for detecting security flaws. They can, for instance, detect large outflows of data from hackers that stand out because users typically only download a small quantity of data that they need.
Some companies are pondering how they can support the ongoing work specific to AI tools, such as juggling the datasets, constructing the models, and then rotating the models to maintain performance.
What can AIops do for security?
While many industries such as AIops focus on practical performance concerns such as how quickly a server responds to a request, others are also using AI algorithms to monitor for the kinds of anomalies that indicate a leak or unauthorized intrusion.
AIops can assist with cybersecurity in many ways. If the website is designed to provide simple, quick answers with at most one user's personal information, then a larger block might suggest a mistake.
AIops may be interested in following these guidelines in the following areas:
- Outflows from servers that dont normally respond or send packets to machines outside the company.
- Atypical SQL queries that are new or rarely seen.
- Atypical requests for encryption keys.
- Responses that are encrypted even though they normally arent or vice versa.
- Unusual load at unusual times. For example, a heavy number of requests in the middle of the night when everyone is normally asleep.
This is especially useful when security breaches are typically quite rare and difficult for a human to detect. An algorithm can monitor thousands of machines and identify the one with a load or behavior that is unusual.
As the workloads shift, AIops algorithms will also evolve. This can be helpful because some attacks require the removal of older software that is no longer used. For example, the models can detect that certain access mechanisms arent used in common use and flag them.
How are big businesses coping with AIops?
All cloud and service providers have regular services for exploring and deploying AI. The initial offerings began relatively simple, but as users began relying on AI algorithms for production tasks, the companies have expanded their services to include maintaining datasets and models as needed.
The most powerful players are also adding new hardware configurations to make it as inexpensive as possible to obtain AI solutions. Some are also creating custom hardware that can speed up processing, often dramatically.
Amazon has developed a custom chip called Inferentia to speed up AI deployments. The process is often performed much more often than training. The Inferentia is claimed to be 70% cheaper than using one of AWS' regular GPU-enabled instances.
IBM has added AIops to its Watson Cloud Pak, so the software supports ongoing AI-based decisions. The tool assists the team monitoring the AI watch for anomalies and adverse events. Intelligent Root Cause Analysis is designed to help the company understand why decisions are made, whether correctly or incorrectly.
Google has developed a line of specialized chips for ML that they call TPUs or Tensor Processing Units, that can provide faster performance and lower costs for AIops. TensorFlow Enterprise is a tool that assists teams who are using TensorFlow open-source software in their production processes.
Microsoft has integrated its AI solutions into many of its products. It's not unusual to discover that the simplest way to interact with AI is as a feature for some of its web tools like Dynamics 365, a business management platform. They're also planning the best solutions for continued delivery of ML solutions with tools like Gandalf, a system that ensures that new models and software are unveiled and maintained.
Nvidia, the world's leading graphics processor manufacturer, supports many cloud options for training and deploying artificial intelligence models through its CUDA architecture. The company continues to support all clouds that are using Nvidia hardware with a collection of tools like Launchpad.
Also read: AIops lessons learned: Be careful when picking a vendor.
What about startups based on AI?
Many ITops and devops businesses offer AI algorithms as well. The same techniques that can detect a failed database or an overloaded server can also detect a problematic server that is executing an AI routine. Good operations tools can solve many problems that artificial intelligence can't.
NewRelic, DataDog, Splunk, PagerDuty, Turbonomic, and DynaTrace are just a few of the most well-known companies that help monitor server and software performance. Their dashboards and other tools are useful for tracking performance.
AIops D is a business aimed at developing microservices that may rely on artificial intelligence to accomplish some of its objectives. The company, founded by Deloitte, also offers consulting services to help develop some suitable microservices to meet business needs. The goal is to develop a set of mostly automated services that manage all of the business processes.
Databricks and DataRobot are forming clouds that collect data and then use the best AI algorithms to construct models. They began as data warehouses or data lakes and evolved to support advanced analysis.
Is there anything that AIops can't do?
Platforms like AIops tackle a wide variety of challenges, but they are only as good as their data. If the data is noisy, incomplete, or full of gaps, the analysis will be less accurate and sometimes completely wrong.
Unusual events can be a challenge. In some cases, AIops platforms are just tasked with flagging unusual events. In these situations, strange patterns that do not match the historical data are easy to spot.
Except in rare instances, the AIops platform is expected to forecast the future. Strange or unusual events may produce misleading results. If the AI model is built from the record and it learns how to behave by studying the past, then a new, unusual event will be something it cannot handle because it has no context or history.
When the AIops platform manages AI models and data gathering, the AIops platform will only support the AI algorithms by making it easier to create new ones. It will not be able to make the algorithms more accurate. AIops can only handle the housekeeping chores.
Next: How Artificial Intelligence (AIops) Can Benefit Businesses