At some point, all old models break and make way for new models. Case in point: Galileo’s observations of Venus in retrograde broke the Earth-at-the-center view of the heavens and compelled Galileo to support heliocentrism, Copernicus’ math-based theory for our solar system. On some level, although not visible to most people outside of IT, the world of IT integrations is as fundamental to the digital era as Earth’s place at the center of the solar system was four centuries ago.
As our IT world expands and our connectedness grows exponentially, another paradigm is collapsing. In Thomas Kuhn’s revolutionary book, The Structure of Scientific Revolutions, Kuhn explores two paths forward. The first, gradualism, is marked by incremental changes that address the current paradigm’s inadequacies. The second is sudden, revolutionary change, like Galileo’s view positing that Earth orbits the sun (which put Galileo on trial for heresy in 1616). Kuhn’s thesis is that revolutionary paradigm shifts succeed when they provide a new ultra-simple and beautiful model that addresses the full system but in a much cleaner way than its predecessor.
Let’s take a moment to visualize the full array of configuration items (CIs) in the IT celestial sky, from mainframe to mobile: routers and switches, gateways and firewalls, servers and storages, databases and applications, microservices and APIs, containers, and cloud services. To keep the lights on, IT operations and monitoring teams monitor everything that happens among these CIs and sub-CIs, and all of their interactions with one another. They integrate with one another so that they can pass data back and forth, all with the objective of playing nicely together to deliver a positive customer experience.
For more than a few decades, our AIOps team has helped IT operations and monitoring teams manage their IT infrastructures by improving their fault management, performance management, and capacity management capabilities. In that time, the team has made scientific observations of these solar systems. In the process, we’ve determined that the old integration-centric model is broken and that it must make way for a new model.
So, you may ask, “What does any of this have to do with Jimi Hendrix and Bob Dylan?”
The IT Orchestra
Believe it or not, Jimi Hendrix plays for the IT orchestra. His partition is network fault management which plays in his style of music at his own speed, in the key of Hendrix.
Bob Dylan’s special talent in the IT orchestra is totally different: his partition is system and application performance management, which plays in his style of music at his own speed, in the key of Dylan.
Mr. Hendrix and Mr. Dylan know very little about each other and don’t play the same kind of music, but each is exceptionally good in his own domain.
By definition, the IT orchestra, like any orchestra, brings a diverse assortment of expert domain managers, such as asset or configuration management DBs, network or system monitoring, fault or performance management systems, identity providers, and application transaction monitoring systems, each representing an IT domain, together in harmony.
Just picture our IT orchestra: Ludwig von Beethoven, Bob Marley, Taylor Swift, Sir Mick Jagger with The Rolling Stones, Freddie Mercury, and Lady Gaga, all playing with Mr. Hendrix and Mr. Dylan. A few other exceptional musicians lurk in the IT shadows, but frankly, we’re not sure what they’re doing.
Synchronizing and aligning all of this diversity are our IT operations and monitoring teams, which are tasked with fault management, performance management, and capacity management to keep our audience (customers) happy and address issues if things go awry.
Mr. Hendrix should be listening to Mr. Dylan, but he’s not. He’s distracted because Sir Mick Jagger is ad-libbing and Ms. Gaga is singing too loud and physically. Mr. Hendrix pauses for a moment to confer with Mr. Dylan to see if they can get on the same page. Ms. Gaga sees this as an opportunity to move to center stage; Bob Marley is pretty OK with everything.
IT integration in the sense of metadata allows an IT orchestra member to interact with another member on a peer-to-peer basis. Mr. Dylan needs to know how to listen to and interpret Mr. Hendrix, and sometimes, vice versa. With so many members of the IT orchestra, that’s a lot of integrations for IT teams to develop, test, and maintain. When new members or special soloists arrive, they require new peer-to-peer integrations. To maintain the flows between all those individual integrations, IT ops and monitoring teams need to grasp these peer-to-peer exchanges to decipher the root cause of fault-, performance- and capacity management issues.
Some of the integrations are handled with APIs, which serve as published guides on how one or more members should communicate with another single peer (for instance, Beethoven APIs inform other members how they should play nicely and communicate with him). Between exacting peer-to-peer integrations and APIs, one would think the IT orchestra would deliver harmonious customer experiences.
When alarms light up the dashboards used by IT operations and monitoring teams, teams struggle to isolate the offending diatonic notes from one of the orchestra members, and identify problems to fix the effect that the error has had on other members. Put in IT terms, when application performance degrades, it takes insightful sleuthing and correlation analysis to diagnose and remediate the problem, all while the performance is underway: Is the root cause in the application? Is it a server or database performance issue? How about the network or firewall? Perhaps it’s a peer-to-peer integration issue?
Integrations do an exceptional job with one-to-one relationships. What they don’t do well—and what is a real problem when you have an orchestra full of diverse domain managers —is one-to-many relationships. While an integration can get Mr. Hendrix and Mr. Dylan playing nicely together, it’s challenging to get both Mr. Hendrix and Mr. Dylan to collaborate with all of the other members of the IT orchestra, even when the house lights go down and the performance is about to start.
APIs typically offer a one-to-many integration. They extract information from one member and load it into one or several other members. Done automatically, this work has great value.
However, APIs have a severe limitation: APIs are built only to communicate, extract and/or load data; they do not transform data into a format suitable to other orchestra members. An API would automatically share Sir Mick Jagger’s lyrics with Beethoven in their raw form. The problem is that Beethoven may not know what he should do with this information or how it may affect his contribution to the orchestra’s performance. Beethoven may a) ignore all or some of the data, b) react incorrectly, c) react correctly, or d) get creative and do something Sir Mick and other orchestra members are not expecting.
Simply put, that’s quite a bit of fun for the IT ops and monitoring teams responsible for delivering a positive customer experience.
But I have good news: A better model is ready to replace this outmoded one.
The goal is to get the IT orchestra members to play nicely together, with less effort, less risk of error, and faster recovery when things go wrong—in short, to deliver an excellent customer experience. We need a new paradigm. The question is, if we have outgrown the old heliocentric approach to IT (aka reliance on peer-to-peer integration and APIs), what model do we put in its place?
The answer: Orchestration.
Middle C: Developing a Common Vernacular
To develop a common vernacular for every player in the orchestra, we start by standardizing and normalizing the structure and values of the metadata database (MDB) for CI and sub-CI information. The example below shows the nomenclature of geolocation metadata for two business units of a company.
|Business Unit A||Business Unit B|
|Metadata Field||Sample Values||Metadata Field||Sample Values|
|Town||Hardin||Post Code||M5J 2N4|
Business Unit A Business Unit B Metadata Field Sample Values Metadata Field Sample Values Continent North America Europe Continent Code NA EU Country Canada US Country Code CN US State Arizona MT Province Code AB MB County BH PR Town Hardin Post Code M5J 2N4
For this global company with multiple business units, the IT celestial sky includes a wide range of CIs and sub-CIs, such as routers and switches, gateways and firewalls, servers and storages, databases and applications, microservices and APIs, containers, and cloud services. These CIs are described by IT operations using metadata. The IT ops team, for example, relies on geographic information to isolate issues by location. It must maintain and reconcile geographic information for all CIs across all business units, an enormous challenge that increases delays and the risk of errors, and adds work.
Standardizing and normalizing the metadata fields and the values contained in those fields is a one-time effort that can be successfully led by corporate IT.
Just as geographic metadata and values must be normalized and standardized following good practices, the same needs to be done for three additional metadata classifications:
- Geolocations (such as Continent, Country, State, City, Site)
- Organizations (such as Company/Organization/Department)
- Services (such as VoIP, Wireless, Billing)
- Technologies (such as Firewalls and Security, Routing, Storage)
When all CI and sub-CI metadata have been standardized and normalized, Mr. Hendrix no longer integrates directly with Mr. Dylan. Instead, every player, from Beethoven to Mick Jagger, integrates with a centralized conductor, who orchestrates the entire orchestra (hence the name). Rather than their music being a product of countless peer-to-peer integrations, which results in a discordant mess, the orchestrated orchestra performs in perfect harmony. All CIs and sub-CIs across the enterprise provide the conductor with normalized and standardized metadata for geolocations, organizations, services and technologies. When the company acquires another company or new business line, for example, the acquired company’s infrastructure is pushed into the same normalized and standardized nomenclature.
Not only does this model for AIOps orchestration lead to more harmonious music, but it also facilitates and accelerates the merger and acquisition process. It also helps the orchestra get back on track if a problem arises. If Mr. Hendrix has a question about how to play nicely with Ms. Gaga, who just joined the orchestra, Mr. Hendrix can get the answer directly from the conductor, the owner, and distributor of all CI and sub-CI metadata. If IT encounters a performance issue, the team can quickly isolate and communicate its source based on the common vernacular defined in the metadata.
For example, if IT uses a wide range of performance, fault and capacity management monitoring tools from various vendors, when one tool raises an alert for application performance, all monitoring tools across all vendors and all CI types instantaneously share and correlate the full horizontal health of the IT service. This gives the IT ops team all the data it needs to identify and fix the issue quickly.
Vertical vs. Horizontal
You may be asking yourself, “What does he mean by “the full horizontal health of the IT service”? Within IT, organizational structures, tooling, performance metrics, dashboards, and reports tend to be oriented to the vertical silos of network, system, and application—fault, performance or capacity management, and so on. In other words, Mr. Hendrix (remember, his instrument is the network fault management domain manager) thinks about network fault only. When the network is playing well, Mr. Hendrix is happy.
But the audience doesn’t care about Mr. Hendrix individually. They are much more interested in how he contributes to the full combined performance of the IT orchestra, called the business services, such as VideoConf and VoIP services. For that reason and based on the ITIL definition, business services are built horizontally across multiple orchestrated IT services; that is, business services cut across the silos until they are provided to the end customer.
Customizing for the Visiting Soloist
But wait. It’s true, the orchestration model described above has no standardized or normalized CI to fit the contingency when Alicia Keys makes an unexpected appearance. That’s why the model accommodates a special guest structure called Custom (as opposed to geolocations, organizations, services, and technologies). By its very nature, this custom structure or group of un-standardized or un-normalized CIs should host a limited and small number of contingencies, a goal that the IT ops team must set for itself—and meet.
Benefits of the AIOps Orchestration Model
Let’s review the invaluable benefits that adopting the new AIOps orchestration model for IT will reap:
- Elimination of unwieldy, awkward, and primordial spreadsheets (aka the poor man’s CMDB) to store CIs, sub-CIs, groups and structures, user and role data, and metadata
- Reliably and consistently high-quality data and metadata
- Precise and accurate root cause analysis across all “guilty” IT networks, systems and applications, CIs and sub-CIs, and all related tickets
- Improved transparency with business views that are inclusive, complete, and oriented to horizontal business services, not vertical IT services
- Fast, efficient and accurate fixes to fault, performance and capacity management issues
- Less disruption when accommodating acquisitions, mergers or other changes
Combined, these benefits will reap the biggest benefits of all: a better understanding of your audience’s experience and better resourcing to improve your user experience.
The musical metaphor and technical contributions were supplied by the AIOps team with Enterprise Studio.
About Enterprise Studio
Enterprise Studio by HCLTech helps organizations make the connections between IT and business that optimize time and multiply value for realizing the full potential of their digital business plans. Our seasoned technologists, coaches, and educators can help you unlock value from existing IT investments to become a stronger, more adaptive organization – in part by leveraging a BizOps approach so that IT outputs are strongly linked to business outcomes.
Whether you’re an established Global 500 company or a new disruptive force in your industry, we can help you navigate the complexities that come with competing in an inter-connected digital era. We are a global solution provider and Tier 1 global value-added reseller of Broadcom CA Technologies and Symantec enterprise software.
Many of our experts at Enterprise Studio are from the former professional services units of CA Technologies and Symantec. For decades, our teams have supported and led organizations to innovation with powerful enterprise software solutions and cutting-edge methodologies – from business and agile management to security, DevOps, AIOps, and automation.