Odock Blog: LLM Infrastructure, Security, and Cost Control

Practical articles for teams building AI products with multiple providers, MCP tools, security guardrails, and production governance requirements.

All articles

Topics

August 19, 2026

Updated

A2A, MCP, and the Agent Standards Race: Why Interoperability Needs Governance

The agent standards race is one of the loudest AI infrastructure stories of 2026. A2A helps agents communicate, MCP connects agents to tools, and enterprises now need a governance layer that controls both.

A2A protocolModel Context ProtocolMCP governanceagent interoperability

Youcef Kaddour

Founder at Odock and AI infrastructure engineer

Read article

What you should take away

1
MCP standardizes agent-to-tool access, while A2A focuses on agent-to-agent communication.
2
The security issue is not the standards themselves; it is unmanaged capability expansion across tools, identities, and agent chains.
3
Odock's MCP-native governance is the right foundation for this shift because it treats tool calls as governed runtime capabilities, not just integrations.

All articles

Clear, technical writing for AI platform teams

AI Governance & Compliance

August 19, 2026

AI Act Transparency and GPAI Enforcement: What Changed After August 2026

The EU AI Act moved from preparation to operational enforcement in August 2026. The practical question for AI teams is now whether they can prove identity, transparency, policy, logging, and oversight for real AI traffic.

Youcef Kaddour

9 min

Read article

Agentic AI

August 19, 2026

From Chatbots to Delegated Work: Why Agentic AI Needs a Runtime Control Plane

The hot AI story of 2026 is not only better models. It is the shift from chat to delegated work. Agents now run longer tasks, call tools, and consume real budgets, which means governance has to move into the runtime path.

Youcef Kaddour

8 min

Read article

AI Gateway

July 18, 2026

From Laptop to Production: Self-Host LLMs with Ollama and vLLM, Then Distribute Them Like a Pro with Odock

You can prototype on Ollama in an afternoon and serve real production traffic on vLLM at 6x the throughput. The part most guides skip is what happens next: turning that self-hosted model into something you can safely hand out to teams, customers, or a pro tier with proper access control and billing.

Youcef Kaddour

12 min

Read article

AI Security

July 17, 2026

AI Agents Are Non-Human Identities Now: The 2026 Credential Sprawl Problem

Your AI agents are not features. They are non-human identities, they hold credentials, they act autonomously, and most security teams have no lifecycle for them at all. Here is why NHI sprawl became 2026's quiet agentic AI risk, and what fixing it actually requires.

Youcef Kaddour

12 min

Read article

AI Governance

July 17, 2026

ISO 42001 vs NIST AI RMF vs EU AI Act: One Gateway, Three Frameworks

ISO 42001, NIST AI RMF, and the EU AI Act are not competing standards, they are three different audiences asking overlapping questions. Procurement wants the certificate, US enterprise wants the risk methodology, the EU wants the legal evidence. Here is how to satisfy all three without building three separate programs.

Youcef Kaddour

11 min

Read article

AI Governance

July 17, 2026

Shadow AI in 2026: Why Banning Tools Fails and Governed Enablement Wins

Two-thirds of employees are already pasting real company data into AI tools nobody approved. The instinct is to block. The data says that fails. Here is the governed-enablement pattern that actually closes the gap, and how a gateway makes it real.

Youcef Kaddour

11 min

Read article