1. The Core Infrastructure Layer (Data & Hosting)

The ServiceNow platform has evolved from a "forms and workflows" engine into an AI-first, data-intensive ecosystem. The most critical modernization is the database layer transformation.

• Modern Standard: RaptorDB (HTAP)
  Insight: A massive shift from legacy MariaDB. RaptorDB allows ServiceNow to handle heavy transactional loads (creating incidents) and complex analytical queries (dashboards) simultaneously without performance degradation.
• Legacy/Standard: MariaDB
  Still used for standard instances, but "Pro" instances and AI-heavy workloads are migrating to RaptorDB.
• Architecture: Multi-Instance
  Unlike traditional SaaS (Multi-Tenant), ServiceNow isolates every customer’s data and runtime into their own unique container set.

2. The Platform Runtime (Backend)

The "engine room" has become more accessible to modern development practices.

• Application Server: Apache Tomcat on Linux
  The core kernel is written in highly optimized, proprietary Java.
• Scripting Engine: Rhino (JavaScript)
  Modern Update: Supports ECMAScript 2021 (ES12) features. Developers can now use arrow functions, let/const, and template literals in server-side scripts.
• Search Technology: AI Search
  Replaces legacy "Zing" text search. It uses machine learning to provide actionable results and GenAI answers directly in the search bar.

3. The User Experience Layer (Frontend)

This layer features the starkest contrast between "Classic" and "Modern" ServiceNow.

• Framework: Now Experience UI Framework (NELF)
  Built on standard Web Components using the Seismic rendering engine.
• Development Tool: UI Builder (UIB)
  A low-code interface for building pages and workspaces.
• Legacy (Phasing Out): AngularJS (v1.x) and Jelly Scripting are being replaced by Configurable Workspaces.

4. The Intelligence Layer (Generative AI)

AI is woven into the stack via the Now Assist family.

• Model: Now LLM
  ServiceNow’s domain-specific Large Language Model combined with external controllers (Azure OpenAI / Gemini).
• Feature: Generative AI Controller
  Allows low-code developers to call LLMs directly from Flow Designer.
• New in 2026: AI Agents
  Autonomous agents that can plan and execute multi-step workflows (e.g., detecting a server outage and drafting a rollback plan) with minimal human intervention.

5. The Workflow & Automation Layer

• Engine: Flow Designer
  The de facto standard replacing Workflow Editor. Runs on a modernized, upgrade-safe Java backend.
• Orchestration: Process Automation Designer (PAD)
  Organizes multiple flows into a linear "Playbook" experience.
• Integration: Integration Hub
  Uses "Spokes" to connect with external systems (Jira, AWS, Azure) without custom scripting.

6. Developer Tools Breakdown: Studio vs. VS Code

With the introduction of the ServiceNow SDK and ServiceNow IDE, the development landscape has split into "Low-Code/Classic" and "Pro-Code/Modern".

Option A: The "Classic" Studio (Browser)

Best for: Quick fixes, admin tasks, and low-code developers.

• Interface: Proprietary tabbed interface inside the browser.
• Strengths: No installation required; visually represents application files; easy access to table creators and form designers.
• Weaknesses: No true IntelliSense; poor Git conflict resolution; cannot work offline; stuck in the "Application File" paradigm rather than true code.

Option B: VS Code + ServiceNow Extension (Desktop)

Best for: Professional developers working on existing scoped apps.

• Interface: Microsoft Visual Studio Code (Desktop).
• Strengths:
  • Sync: Bi-directional sync (save in VS Code, updates in instance).
  • TypeScript Support: Full support for modern JS/TS syntax with the new SDK.
  • Tools: seamless integration with GitHub Copilot for code generation.
• Key Feature: "Offline" capability. You can write code on a flight, and sync it when you reconnect.

Option C: ServiceNow IDE (New in Zurich)

Best for: Building complex apps using the new "Fluent" language.

• Description: A full instance of VS Code running inside the ServiceNow browser tab.
• The "Fluent" Paradigm: Instead of creating XML records for metadata, you write TypeScript code to define tables, ACLs, and Business Rules.
  Example: createTable('x_app_todo').addColumn('status', 'string').
• Why it matters: This allows "Infrastructure as Code" (IaC) for ServiceNow. You can review metadata changes in a Pull Request just like standard software engineering.

Summary: Legacy vs. Modern Stack

Strategic Insight for Developers

The "Golden Path" for modern solution design is:

1. Store data in standard tables (leveraging RaptorDB).
2. Build logic in Flow Designer (low-code first) or Fluent/TypeScript (pro-code).
3. Create interfaces in UI Builder (Workspaces).
4. Enhance with Now Assist (GenAI) for automation.