Silhouette of Business team show arm up and flag on top of the mountain.
PHOTO: Adobe

How many "Ops" can you fit into the enterprise? A limitless number it would seem and many of them under the XOps umbrella. The goal of XOps, which includes DataOps, MLOps, ModelOps, and PlatformOps, is to create an enterprise technology stack that enables automation and reduces the duplication of technology and processes. In simple terms it is the natural evolution of DataOps in a workplace, and across the enterprise, to enable AI and Machine Learning (ML) workflows.

There is also one other ingredient that makes it a key element of the digital workplace. XOps enables data and analytics processionals to operationalize their processes and automation from the beginning rather than to address this issue as an afterthought. In this case, according to the Dev/Central community to "operationalize” software means to apply a systematic approach to automating software building and orchestrating operational processes in a way that meets measurable, defined goals that align with business priorities.

Why XOps?

So why is this even a "thing" now? First, because it has managed to creep onto Gartner’s Top 10 Trends list for data and analytics, which means organizations are starting to seriously look at it as a workplace option, and second, as data leaders are now directly implicated in digital transformation initiatives, it is becoming part of many digital transformation processes in respect to AI and analytics.

To understand XOps you need to put it in the context of DevOps and the way that evolved into DataOps as it applies to workflows.  DevOps is a set of practices that combines software development (Dev) and IT operations (Ops). It aims to shorten systems' development life cycles and provide continuous delivery with high software quality. DevOps is complementary to agile software development while several DevOps aspects have become part of the Agile methodology.

There are four key metrics that are key to assessing the effectives of DevOps according to Google Cloud:

  • Deployment Frequency: How often an organization successfully releases to production
  • Lead Time for Changes: The amount of time it takes a commit to get into production
  • Change Failure Rate: The percentage of deployments causing a failure in production
  • Time to Restore Service: How long it takes an organization to recover from a failure in production.

The application of continuous delivery and DevOps to data analytics has been termed DataOps. DataOps tries to integrate data engineering, data integration, data quality, data security, and data privacy with operations. It applies principles from DevOps, agile development and the statistical process control, to improve the cycle time of extracting value from data analytics.

Even if this is all XOps was, it’s easy to see its role in the digital workplace. However, it contains much more. With the growth of all the different Ops disciplines, XOps looks to accelerate processes and improve the quality of what is being delivered in:

  • Software (DevOps)
  • Data (DataOps)
  • AI models (MLOps)

Gartner explains the situation like this, the multiplication of Ops disciplines stemming out of DevOps best practices has caused significant confusion in the marketplace. Yet, their reconciliation can bring significant advantages to organizations that are able to harmonize those disciplines.” 

It adds that XOps practices link development, deployment and maintenance together to create a shared understanding of requirements, transfer of skills and processes for monitoring and maintaining analytics and AI artifacts.

In sum XOps is about make your data and analytics deployments work better and in conjunction with other software disciplines.

Related Article: Why Artificial Intelligence May Not Offer the Business Value You Think

The Emergence of XOps

So how did this XOps situation emerge? Chris Bergh, CEO of Cambridge, Mass.-based DataKitchen, a DataOps consultancy and platform provider that manages analytics creation and operations, points out that in the past couple of years, there has been a tremendous proliferation of acronyms with the Ops suffix. This was started in the software space by the merger of development (dev) and IT operations (Ops). Since then people have been creating new Ops terms at a pretty rapid pace. DevOps and DataOps serve as the foundation for all other Ops.

In the software industry, companies began to understand that lean manufacturing principles could be equally transformative in the context of software development. Lean seeks to identify waste in manufacturing processes by focusing on eliminating errors, cycle time, collaboration, and measurement. Lean is about self-reflection and seeking smarter, less wasteful dynamic solutions together. Combined with Agile, DevOps has helped many companies attain a leadership position in their markets.

“More recently, data analytics organizations are applying lean principles to their methodologies. Enterprises following this path find that these methods help data science/engineering/BI/governance teams produce better results more efficiently,” he explained.

However, DevOps by itself is insufficient for data organizations. Data organizations have more complex technical environments, as well as two pipelines — development and production. As a result, DataOps (including its derivatives ModelOps, MLOps, Analytic Ops, DataGovOps) emerged and are tuned to the special needs of data analytics teams. DataOps companies run as smoothly as oiled data factories, enabling teams to quickly iterate and innovate new analytics These companies are gaining a significant competitive advantage because they can deliver high-quality on-demand analytics more rapidly than their peers.

Unifying Your Ops

The complexity of AI is driving growing interest in XOps, and more specifically ModelOps, Carlos Melendez, COO of Palo Alto, Calif.-based Wovenware, a nearshore provider of software engineering and AI service, added. As AI continues to proliferate, so too does the number of algorithms deployed to address specific business problems, so organizations must deploy multiple algorithms to attack new problems.  But how can companies effectively initiate ModelOps and ensure it becomes a key methodology for AI development and deployment?  Key steps include:

  • Removing the organizational siloes to enable greater collaboration between software engineers, data scientists and IT staff
  • Establishing the rules of operation and automation to standardize core processes and automation tools
  • Monitoring KPIs to determine its effectiveness and its consistent training and growth

XOps is an idea developed for one overarching purpose: the productionalization of AI. The pace of digitization with machine learning has allowed companies to better automate AI adoption into their development processes.

 “Yet it's becoming less of a cutting-edge bonus and more of a necessity, as the time needed to implement AI in a siloed landscape leads to increasingly-lengthy project delays. Building an encompassing XOps process allows automation to be built directly into the data pipeline so there's a seamless transition to what was previously called the ModelOps phase,” he said.

“This kind of integration is meant to reduce the processing times that previously caused bottlenecks in production and made upgrades take months or years. Essentially, it's the unifying of various Ops best practices under the AI umbrella.”

Cross Department Collaboration

As development models become more advanced to deal with the different needs of each company, it's not unusual to see new names come up for different production methods.

At its core, Phil Strazzulla of Cambirdge, Mass.-based software researchers Select Software Reviews said, XOps is a development infrastructure that follows the commonly used practices established by DevOps. The benefit of this infrastructure is that it inherently supports the collaboration of different departments. This methodology helps organizations scale as they and their projects grow. Since XOps includes different departments and workflows under its umbrella, it assists in the collaboration between units. This helps breaks down silos within a company and supports efficiency.

 “Overall, XOps reinforces the idea that different development teams are cross-functional and work with one another. XOps seems to be the adoption of Agile principles in a DevOps environment, "which helps a company scale and produce efficiently. This infrastructure may help businesses formalize processes they're already engaging in to improve the scalability and automation of their workflows,” he said.