One of many ideas in our upcoming e-book Structure as Code is the flexibility for architects to design automated governance checks for necessary architectural issues, creating quick suggestions loops when issues go awry. This concept isn’t new—Neal and his coauthors Rebecca Parsons and Patrick Kua espoused this concept again in 2017 within the first version of Constructing Evolutionary Architectures, and lots of of our purchasers adopted these practices with nice success. Nevertheless, our most formidable objectives have been largely thwarted by a typical downside in trendy architectures: brittleness. Thankfully, the appearance of the Mannequin Context Protocol (MCP) and agentic AI have largely solved this downside for enterprise architects.
Health Capabilities
Constructing Evolutionary Architectures defines the idea of an architectural health operate: any mechanism that gives an goal integrity test for architectural traits. Architects can consider health capabilities type of like unit exams, however for architectural issues.
Whereas many health capabilities run like unit exams to check construction (utilizing instruments like ArchUnit, NetArchTest, PyTestArch, arch-go, and so forth), architects can write health capabilities to validate all kinds of necessary checks…like duties usually reserved for relational databases.
Health capabilities and referential integrity
Think about the structure illustrated in Determine 1.
In Determine 1, the staff has determined to separate the information into two databases for higher scalability and availability. Nevertheless, the widespread drawback of that method lies with the truth that the staff can now not depend on the database to implement referential integrity. On this scenario, every ticket will need to have a corresponding buyer to mannequin this workflow appropriately.
Whereas many groups appear to suppose that referential integrity is barely potential inside a relational database, we separate the governance exercise (information integrity) from the implementation (the relational database) and understand we will create our personal test utilizing an architectural health operate, as proven in Determine 2.
In Determine 2, the architect has created a small health operate that displays the queue between buyer and ticket. When the queue depth drops to zero (which means that the system isn’t processing any messages), the health operate creates a set of buyer keys from the buyer service and a set of buyer international keys from the ticket service and asserts that the entire ticket international keys are contained throughout the set of buyer keys.
Why not simply question the databases immediately from the health operate? Abstracting them as units permits flexibility—querying throughout databases on a relentless foundation introduces overhead that will have unfavourable negative effects. Abstracting the health operate test from the mechanics of how the information is saved to an summary information construction has not less than a few benefits. First, utilizing units permits architects to cache nonvolatile information (like buyer keys), avoiding fixed querying of the database. Many options exist for write-through caches within the uncommon occasion we do add a buyer. Second, utilizing units of keys abstracts us from precise information gadgets. Information engineers choose artificial keys to utilizing area information; the identical is true for architects. Whereas the database schema would possibly change over time, the staff will at all times want the connection between clients and tickets, which this health operate validates in an summary means.
Who executes this code? As this downside is typical in distributed architectures akin to microservices, the widespread place to execute this governance code is throughout the service mesh of the microservices structure. Service mesh is a normal sample for dealing with operational issues in microservices, akin to logging, monitoring, naming, service discovery, and different nondomain issues. In mature microservices ecosystems, the service mesh additionally acts as a governance mesh, making use of health capabilities and different guidelines at runtime.
This can be a widespread means that architects on the utility stage can validate information integrity, and we’ve applied these kind of health capabilities on a whole bunch of tasks. Nevertheless, the specificity of the implementation particulars makes it troublesome to broaden the scope of these kind of health capabilities to the enterprise architect stage as a result of they embody too many implementation particulars about how the venture works.
Brittleness for metadomains
One of many key classes from domain-driven design was the thought of retaining implementation particulars as tightly certain as potential, utilizing anticorruption layers to forestall integration factors from understanding too many particulars. Architects have embraced this philosophy in architectures like microservices.
But we see the identical downside right here on the metalevel, the place enterprise architects want to broadly management issues like information integrity but are hampered by the space and specificity of the governance requirement. Distance refers back to the scope of the exercise. Whereas utility and integration architects have a slim scope of accountability, enterprise architects by their nature sit on the enterprise stage. Thus, for an enterprise architect to implement governance akin to referential integrity requires them to know too many particular particulars about how the staff has applied the venture.
Certainly one of our greatest international purchasers has a job inside their enterprise structure group known as evolutionary architect, whose job is to determine international governance issues, and we now have different purchasers who’ve tried to implement this stage of holistic governance with their enterprise architects. Nevertheless, the brittleness defeats these efforts: As quickly because the staff wants to alter an implementation element, the health operate breaks. Despite the fact that we frequently sofa health capabilities as “unit exams for structure,” in actuality, they break a lot much less typically than unit exams. (How typically do adjustments have an effect on some basic architectural concern versus a change to the area?) Nevertheless, by exposing implementation particulars exterior the venture to enterprise architects, these health capabilities do break sufficient to restrict their worth.
We’ve tried quite a lot of anticorruption layers for metaconcerns, however generative AI and MCP have supplied the perfect resolution up to now.
MCP and Agentic Governance
MCP defines a normal integration layer for brokers to question and eat capabilities inside a selected metascope. For instance, groups can arrange an MCP server on the utility or integration structure stage to reveal instruments and information sources to AI brokers. This gives the right anticorruption layer for enterprise architects to state the intent of governance with out counting on implementation particulars.
This permits groups to implement the kind of governance that the strategically minded enterprise architects need however create a stage of indirection for the small print. For instance, see the up to date referential integrity test illustrated in Determine 3.
In Determine 3, the enterprise architect points the overall request to validate referential integrity to the MCP server for the venture. It in flip exposes health capabilities by way of instruments (or information sources akin to log recordsdata) to hold out the request.
By creating an anticorruption layer between the venture particulars and enterprise architect, we will use MCP to deal with implementation particulars in order that when the venture evolves sooner or later, it doesn’t break the governance due to brittleness, as proven in Determine 4.
In Determine 4, the enterprise architect concern (validate referential integrity) hasn’t modified, however the venture particulars have. The staff added one other service for consultants, who work on tickets, which means we now must validate integrity throughout three databases. The staff adjustments the interior MCP software that implements the health operate, and the enterprise architect request stays the identical.
This permits enterprise architects to successfully state governance intent with out diving into implementation particulars, eradicating the brittleness of far-reaching health capabilities and enabling way more proactive holistic governance by architects in any respect ranges.
Defining the Intersections of Structure
In Structure as Code, we talk about 9 completely different intersections with software program structure and different elements of the software program growth ecosystem (information representing certainly one of them), all expressed as architectural health capabilities (the “code” a part of structure as code). In defining the intersection of structure and enterprise architect, we will use MCP and brokers to state intent holistically, deferring the precise particulars to particular person tasks and ecosystems. This solves one of many nagging issues for enterprise architects who wish to construct extra automated suggestions loops inside their methods.
MCP is nearly ideally suited to this function, designed to reveal instruments, information sources, and immediate libraries to exterior contexts exterior a selected venture area. This permits enterprise architects to holistically outline broad intent and depart it to groups to implement (and evolve) their options.
X as code (the place X may be all kinds of issues) usually arises when the software program growth ecosystem reaches a sure stage of maturity and automation. Groups tried for years to make infrastructure as code work, however it didn’t till instruments akin to Puppet and Chef got here alongside that would allow that functionality. The identical is true with different “as code” initiatives (safety, coverage, and so forth): The ecosystem wants to supply instruments and frameworks to permit it to work. Now, with the mixture of highly effective health operate libraries for all kinds of platforms and ecosystem improvements akin to MCP and agentic AI, structure itself has sufficient help to affix the “as code” communities.
Study extra about how AI is reshaping enterprise structure on the Software program Structure Superstream on December 9. Be part of host Neal Ford and a lineup of consultants together with Metro Financial institution’s Anjali Jain and Philip O’Shaughnessy, Vercel’s Dom Sipowicz, Intel’s Brian Rogers, Microsoft’s Ron Abellera, and Equal Consultants’ Lewis Crawford to listen to hard-won insights about constructing adaptive, AI-ready architectures that help steady innovation, guarantee governance and safety, and align seamlessly with enterprise objectives.
O’Reilly members can register right here. Not a member? Join a 10-day free trial earlier than the occasion to attend—and discover all the opposite sources on O’Reilly.
