--- title: "Multi Format Repositories | Cloudsmith" description: "A quick look at how Cloudsmith repositories are fully Multi Format, and how you can mix and match different package types in one repository." canonical_url: "https://cloudsmith.com/blog/multi-format-repositories" last_updated: "2021-08-06T00:00:00.000Z" --- # Multi Format Repositories | Cloudsmith [Dependency confusion](https://cloudsmith.com/blog/dependency-confusion-attacks) is a software supply chain weakness that arises from how package managers resolve dependencies across multiple sources. It does not rely on complex exploits. It relies on normal dependency resolution. If a build resolves from both private and public registries, and a public package shares the same name as an internal dependency, the public package may be selected. That is sufficient to introduce unintended code into a build. Modern dependency trees often include extensive transitive dependencies, increasing the number of automatic resolution decisions made during builds. The issue was [widely documented in 2021 ](https://medium.com/@alex.birsan/dependency-confusion-4a5d60fec610)when researchers demonstrated it against more than 35 large organizations by publishing public packages that matched internal names. Since then, dependency confusion has remained a known and preventable risk in modern build systems. ## Where risk appears in practice Modern builds commonly blend: - Internal packages - Public open source dependencies - Upstream proxies - Cached artifacts If resolution rules are not explicit, package managers make decisions on your behalf. Different ecosystems provide different structural protections and different pitfalls. ### Ecosystem specific considerations #### Python PyPI operates on a flat global namespace. There are no built-in organizational scopes and no domain bound identifiers. Package names are globally unique and first come, first served. This means internal package names, for example acme utils or internal ml core, are structurally indistinguishable from public ones unless additional controls are applied. Python projects frequently configure multiple indexes. It is common to see: - Internal private mirrors - Vendor-hosted repositories such as PyTorch wheel indexes for CUDA builds - Regional mirrors - Index url combined with extra index url Python tooling such as pip, uv, and Poetry evaluates candidate versions across configured sources. If the same package name exists in more than one index, the resolver may select the highest compatible version, regardless of origin. Real-world incidents demonstrate this behavior. [In 2022, a malicious package mimicking a PyTorch dependency](https://www.reversinglabs.com/blog/pytorch-supply-chain-attack-dependency-confusion-burns-devops) was published to PyPI and installed via pip resolution during nightly builds, leading to data exfiltration before detection. Without explicit trust boundaries or source mapping, this multi index behavior introduces ambiguity. If an internal package name is exposed and a higher version appears on a public index included in resolution, the public package may be selected. Python’s flat namespace model, combined with common multi index configurations, makes careful source control particularly important. #### npm npm supports scoped packages using the @org/package format. Once registered, a scope is controlled by that organization. However, unregistered scopes remain claimable. If an organization uses scoped internal packages but has not reserved the scope publicly, an attacker could register it and publish similarly named packages, for example @org/app. A configuration mistake in development or CI could then result in the malicious package being resolved and executed, including via lifecycle hooks such as preinstall. Unscoped internal package names carry exposure similar to Python. #### Maven Maven Central uses domain-based groupId verification, providing stronger ownership guarantees when correctly configured. However, inconsistent repository usage or additional repositories can reintroduce ambiguity in resolution order if trust boundaries are not clearly defined. #### NuGet While NuGet supports ID prefix reservation to prevent public name squatting, resolution ambiguity can still arise if multiple feeds are configured. **Key safeguards include:** - ID prefix reservationReserve your organization’s package ID prefix on nuget.org to prevent public impersonation. - Package Source Mapping Use NuGet 6+ source mapping to bind specific package IDs or prefixes to a single trusted feed. - Strict nuget.config configurationClear default feeds and explicitly define approved sources to prevent unintended restores. - Signature enforcementNuGet repository signing using X.509 certificates. When signature verification is enforced in the NuGet or .NET CLI, consumers can verify that a package originated from the expected repository and has not been tampered with. #### Docker Docker Hub differs from language ecosystems because image references are typically registry qualified, making classic multi-source confusion less common. However, **related risks remain:** - Implicit Docker Hub fallback when no registry is specified - Unregistered namespace squatting - Mutable tags such as latest These are mitigated through fully qualified image references, namespace reservation, digest pinning using SHA256, and image signing. ## A layered mitigation approach Preventing dependency confusion does not require dramatic changes to developer workflows. It requires clear trust boundaries in how dependencies are resolved. In practice, this is implemented through an internal artifact repository that controls how dependencies are proxied, cached, and resolved. Key elements include: - Centralizing dependency resolution - Defining explicit upstream trust - Controlling namespace ownership - Using lockfiles and reproducible builds - Enforcing artifact signing and verification ### 1. Central artifact repository Route all builds through a central artifact repository that acts as the single source of truth for dependencies. Public packages should be proxied and cached internally rather than resolved directly from the internet. ### 2. Define upstream trust Explicitly distinguish between trusted and untrusted sources. Cloudsmith provides upstream trust controls that allow repositories to designate upstream sources accordingly and control how packages are blended during resolution. If a package exists in a trusted source, it cannot be overridden by an untrusted one. Cloudsmith’s Upstream Trust currently supports: - Python - Maven - npm Support for additional ecosystems is expanding. In addition to source trust controls, policy enforcement provides another layer of protection. [Cloudsmith’s Enterprise Policy Manager](https://cloudsmith.com/product/enterprise-policy-manager) allows organizations to define rules around what packages are permitted, including conditions based on vulnerability data, malware data, version constraints, or upstream origin. This enables teams to combine resolution boundaries with policy-based controls. ### 3. Reserve namespaces Where supported: - Use npm scopes - Reserve NuGet ID prefixes - Use domain based Maven group IDs Namespace ownership significantly reduces collision risk. ### 4. Use lockfiles Lockfiles improve determinism, traceability, and visibility into unexpected changes. They do not eliminate dependency confusion, but they reduce instability and make tampering easier to detect. ### 5. Require signing Signing adds cryptographic integrity to artifacts. Cloudsmith signs hosted packages by default. Cloudsmith supports both native and non-native signing across package formats, including native signing for [Docker, NuGet, and Swift](https://cloudsmith.com/blog/native-signing-support-in-cloudsmith-extended-to-docker-nuget-and-swift). Signature validation workflows in supported ecosystems provide an additional verification layer if substitution is attempted. Signing strengthens integrity guarantees but does not replace resolution controls. ## Conclusion Dependency confusion is not a sophisticated exploit. It is a consequence of ambiguous resolution behavior across multiple sources. Preventing it requires: - A central artifact repository serving as the single source of truth for dependencies - Explicit upstream trust boundaries - Namespace ownership - Lockfiles and reproducible builds - Artifact signing and verification Clear boundaries remove ambiguity. Removing ambiguity removes the attack path. Stop leaving your build integrity to chance. [**Book a demo**](https://cloudsmith.com/book-a-demo) today to see how Cloudsmith's enterprise-grade repository controls can remove ambiguity and harden your delivery process. **TL;DR:** LLMOps is the operational framework for managing the lifecycle of Large Language Models (LLMs). Unlike DevOps, which focuses on deterministic code, **LLMOps artifact management** must handle probabilistic assets like prompts, embeddings, and fine-tuned models. This shift requires a move from standard CI/CD to specialized **LLM pipeline management** to ensure system traceability and trust. ## What is LLMOps? **LLMOps (Large Language Model Operations)** is a specialized set of practices for automating and managing the end-to-end lifecycle of LLM-powered applications. It extends MLOps principles to address the unique requirements of generative AI, specifically focusing on **LLM lifecycle management**, prompt engineering, and vector-based data flows. While DevOps focuses on application code and MLOps on traditional machine learning models, **LLMOps** handles the massive complexity of: - **Foundation and fine-tuned models:** Managing base models and their task-specific variants. - **Prompt artifacts:** Versioning the system instructions that dictate model behavior. - **Embeddings and vector indexes:** Curating the "knowledge" used in Retrieval-Augmented Generation (RAG) systems. - **Dynamic inference behavior:** Monitoring outputs that change even when input code remains the same. In essence, LLMOps is about operationalizing AI rather than just software binaries. ## LLMOps vs DevOps: Why the difference matters The debate of **LLMOps vs DevOps** isn't about choosing one over the other; it’s about understanding where **DevOps tooling limitations for AI** begin. DevOps is built for deterministic systems; if you deploy the same code, you get the same result. LLM pipelines are probabilistic, meaning the same "code" (prompt) can yield different outputs. ```json { "_key": "5a4b45476f69", "_type": "tableBlock", "caption": "LLMOps vs DevOps", "firstRowIsHeader": true, "markDefs": null, "table": { "rows": [ { "_key": "d673a428-3c6a-49c3-87fa-b7dfdff2ec98", "_type": "tableRow", "cells": [ "Category", "DevOps", "LLMOps" ] }, { "_key": "453e4f7a-37ed-4c9c-8bd9-1610932f52de", "_type": "tableRow", "cells": [ "Primary focus", "Application code and services", "Large language models and AI systems" ] }, { "_key": "8bef99c7-76e1-4dc0-81f3-8f96bc8be683", "_type": "tableRow", "cells": [ "Pipeline type", "Linear CI/CD pipelines", "LLM pipelines (training, fine-tuning, evaluation)" ] }, { "_key": "30efc20f-d752-4f94-b718-b4bac0e30ace", "_type": "tableRow", "cells": [ "Artifact types", "Software artifacts (containers, binaries)", "AI artifacts (models, prompts, embeddings)" ] }, { "_key": "30be9c5f-912a-4b10-b7ba-8a3d3b594016", "_type": "tableRow", "cells": [ "Behavior", "Deterministic and reproducible", "Probabilistic and context-dependent" ] }, { "_key": "17362cff-191b-4f7a-9c4a-3c2b2a8be477", "_type": "tableRow", "cells": [ "Change frequency", "Deliberate versioning", "Rapid iteration of prompts and datasets" ] }, { "_key": "b1fc02b1-3d99-4d38-aa79-6a8215086984", "_type": "tableRow", "cells": [ "Traceability", "Moderate (log-based)", "Critical (lineage-based for compliance)" ] } ] } } ``` The core takeaway is that the shift from **DevOps artifact management** to **AI artifact management** involves handling much larger, more volatile assets that directly influence the "logic" of the application. ## Why artifact management matters in LLMOps In a traditional app, an artifact is just a compiled file. In AI, **artifacts are the system.** Without robust **artifact management for LLMs**, teams face a "black box" problem where they cannot explain why a model suddenly began hallucinating or failing. **Effective AI artifact management solves for:** - **Reproducibility:** Re-creating a specific model state using exact dataset snapshots. - **Auditability:** Tracking the lineage of a prompt to meet emerging AI regulations. - **Rollback safety:** Quickly reverting to a previous "known good" version of a prompt or embedding index. - **Cost efficiency:** Preventing redundant training by reusing existing **model artifacts**. ## What artifacts do LLM pipelines produce? Modern **LLM pipeline management** generates a diverse array of non-code assets across the **AI model lifecycle**. Understanding these is key to moving beyond simple script-based deployments. #### Common LLM artifacts: - **Model artifacts:** These include base foundation models (like Llama 3 or GPT-4), fine-tuned adapters (LoRA/QLoRA), and quantized versions for edge deployment. - **Dataset versioning:** Snapshots of training data, evaluation sets (Golden Sets), and synthetic data used for testing. - **Prompt artifacts:** Versioned system prompts, few-shot examples, and complex prompt chains that function as the "new source code." - **Embeddings management:** Vector database snapshots and the specific embedding models (e.g., Ada, BERT) used to generate them. - **Inference artifacts:** Production logs, "LLM-as-a-judge" evaluation scores, and human-in-the-loop feedback. ## MLOps vs LLMOps: Where traditional approaches fall short Many teams assume their existing MLOps stacks can handle LLMs. However, **MLOps vs LLMOps** highlights a critical gap: **prompt versioning.** Traditional MLOps tools aren't built to treat a 50-word text string (a prompt) as a deployment-critical artifact. Furthermore, the **inference artifacts** in LLMOps are much richer, requiring semantic monitoring rather than just simple accuracy metrics. #### Feature store vs Artifact repository A common point of confusion is the choice between a **feature store vs artifact repository**: - **Feature stores** are for structured data used in tabular ML. - **Artifact repositories** (like weights and biases or MLflow) are the "System of Record" for the unstructured models and prompts that define an LLM app. ## Challenges and best practices for LLMOps Managing these assets comes with significant **challenges of artifact management in LLMOps**, including massive file sizes and the high velocity of prompt changes. #### LLMOps best practices: - **Treat prompts as code:** Store prompts in version-controlled repositories, not hardcoded in your app. - **Centralize your artifact registry:** Use a single source of truth for all models and embeddings to avoid "shadow AI" across teams. - **Automate lineage tracking:** Ensure every inference result is traceable back to the specific model version, prompt, and dataset used. - **Implement evaluation gates:** In your **LLM workflows**, never promote an artifact to production without passing an automated evaluation suite. ## FAQ: Frequently asked questions on LLMOps - ### How is LLMOps different from DevOps? LLMOps manages probabilistic AI assets like models and prompts, while DevOps manages deterministic code and binaries. LLMOps requires specialized pipelines for evaluation and fine-tuning that don't exist in traditional CI/CD. - ### Why does artifact management matter in LLMOps? It ensures that every AI output is traceable and reproducible. Without it, you cannot debug hallucinations, comply with AI audits, or reliably roll back failed updates. - ### What are the most important LLMOps workflows? Key workflows include data ingestion for RAG, automated prompt evaluation, model fine-tuning, and continuous monitoring of inference quality. ## Final thoughts The future of software is no longer just about code; it’s about **artifacts, intelligence, and trust.** As LLMs move from experiments to core infrastructure, the transition from DevOps to LLMOps is inevitable. Teams that master **artifact management for LLMs** today will be the ones building the most reliable, scalable, and auditable AI systems of tomorrow. To manage LLMOps at enterprise scale, use Cloudsmith as your single source of truth. Discover how by [booking your free demo](https://cloudsmith.com/book-a-demo) today. **Cloud migration** is rarely just an infrastructure move. For most DevOps and platform teams in 2026, it’s a once-in-a-decade opportunity to rethink tooling, eliminate legacy bottlenecks, and modernize the [**software supply chain**](https://cloudsmith.com/product/software-supply-chain-security) end-to-end. One of the most overlooked, but highest-impact, areas to revisit during this transition is **artifact management**. As organizations shift workloads and security controls into the cloud, the limitations of [**legacy artifact repositories**](https://cloudsmith.com/blog/the-true-cost-of-legacy-artifact-management) quickly become visible. What once worked in on-premise environments often creates friction, risk, and massive operational overhead in a cloud-native world. This is why **cloud migration** isn’t just a "lift-and-shift" event. It’s the ideal moment to reassess how you store, secure, and distribute artifacts across modern **cloud migration DevOps** workflows. ## Why legacy artifact repositories struggle during cloud migration Traditional, self-hosted repositories were designed for static infrastructure and perimeter-based security. Cloud environments invert those assumptions. During migration, teams commonly encounter: - **Scaling constraints:** Legacy tools often require manual server provisioning or expensive over-capacity planning. - **Operational toil:** Managing patches, database tuning, and storage maintenance for your own repository steals focus from your core product. - **Performance bottlenecks:** When global teams depend on a single on-premise instance, latency sabotages developer velocity. Keeping a legacy repository while modernizing everything else often results in a "partially modern" stack with legacy risk still embedded in your delivery pipeline. ## Cloud migration exposes hidden software supply chain risk Modern cloud adoption increases velocity, but speed without control amplifies risk across the **software supply chain**. Common exposure points include: - **Unverified third-party dependencies:** Cloud-scale builds pull in thousands of external packages that need immediate scanning. - **Inconsistent provenance:** Difficulty tracking exactly "who built what and where" across fragmented environments. - **Limited policy enforcement:** Brittle legacy controls that can't handle the dynamic nature of cloud-native deployments. Re-evaluating artifact management during migration allows teams to embed **zero-trust** governance exactly when redesigning their pipelines. ## The modernization opportunity: Fully managed artifact repositories Cloud migration creates the perfect window to replace self-hosted infrastructure with a fully managed** artifact repository** built for elasticity and global distribution. Modern platforms like **Cloudsmith** deliver: - **Infinite scalability:** No more storage planning or maintenance; the platform automatically scales with your builds. - **Edge performance:** A built-in Package Delivery Network (PDN) delivers artifacts worldwide to reduce latency. - **Integrated security:** Features like automated vulnerability scanning and signature verification are baked in, not bolted on. Instead of recreating legacy architecture in the cloud, organizations can move directly to a fully managed model aligned with cloud‑native principles. This shift transforms artifact management from a maintenance task into a strategic layer of the delivery platform. ## When to move: Aligning your migration strategy Teams often postpone modernization because migration feels complex. However, delaying the decision typically leads to "double migration" work, migrating the legacy tool today and replacing it tomorrow. **Aligning modernization with your cloud move avoids:** 1. **Re-architecting pipelines twice:** Design your CI/CD for your final destination, not a temporary stop. 2. **Moving massive stores twice:** Cloud-native migration scripts (like the Cloudsmith CLI) handle the transfer of binaries and metadata once. 3. **Carrying legacy debt:** Ensure your new cloud environment launches with a clean, high-performance foundation from day one. ## Signs your artifact management is holding you back Before moving to the cloud, audit your current state. If these "silent killers" sound familiar, a lift-and-shift solution will only migrate your technical debt: - **Manual maintenance toil:** Your team spends hours every month on repository upgrades, patching, and storage "garbage collection". - **The "slow download" tax:** Global developers or remote build agents face high latency because your on-premise repository lacks a global distribution network. - **Compliance blind spots:** You struggle to provide a complete "bill of materials" (SBOM) or audit trail for a security incident. - **Brittle CI/CD scripts:** Your pipelines rely on custom, "home-grown" scripts to move packages between environments because your tool doesn't support native promotion workflows. ## The strategic ROI: What actually changes? Modernizing your **artifact repository** during a cloud move isn't just a technical swap; it delivers measurable business impact: - **Developer velocity:** By eliminating manual bottlenecks and enabling faster access to dependencies, teams often see **43% faster release cycles**. - **Zero-trust security:** Centralized policy enforcement and automated vulnerability scanning move security from a "final check" to an integrated part of the build. - **Operational efficiency:** Moving to a fully managed** artifact repository** removes the "toil" of server management, allowing your DevOps engineers to focus on product innovation rather than infra-maintenance. - **Total cost of ownership (TCO):** You trade hidden infrastructure costs and administrative salaries for a predictable, transparent, fully managed model. ## Evaluating alternatives: Beyond JFrog and Nexus Many organizations begin cloud migration using legacy tools like **JFrog Artifactory** or **Sonatype Nexus**, only to find they were built for a different era of infrastructure. Modern cloud-native platforms eliminate the need to manage repository infrastructure while delivering stronger governance and global performance. As a result, more teams are looking for [**JFrog alternatives**](https://cloudsmith.com/switch/jfrog-artifactory) and [**Nexus alternatives**](https://cloudsmith.com/switch/sonatype-nexus) that offer a fully managed, "Zero-Ops" experience. For teams ready to [**migrate, Cloudsmith**](https://docs.cloudsmith.com/migrating-to-cloudsmith?_gl=1*1tuvjc8*_gcl_au*MjA0MTMxOTg3MC4xNzYzMzg3ODg3*_ga*NDcwNDc1ODMuMTcyMzAyNjg2MQ..*_ga_6KCWZ6W3Y9*czE3NzAyOTA3NDEkbzE1MiRnMSR0MTc3MDI5MTAwMiRqNDEkbDAkaDA.*_ga_H5NBQJ0NGM*czE3NzAyOTA3NDEkbzMxMiRnMSR0MTc3MDI5MTAwMiRqNDAkbDAkaDkzNTk0MDIyMw..) streamlines the process to minimize disruption and accelerate value realization. ## Why global leaders migrate artifact management to Cloudsmith As teams evaluate **alternatives to JFrog and Nexus**, they increasingly move to **Cloudsmith**, designed specifically for modern DevOps and secure software delivery for the security landscape of 2026 and beyond. **The Cloudsmith advantage:** - **Zero-ops architecture:** A true cloud-native, fully managed experience with no databases to manage and no servers to patch. - **Built-in package delivery network (PDN):** Hundreds of nodes deliver artifacts from the edge, ensuring your global build agents always have high-speed access. - **Universal format support:** One single source of truth for Docker, npm, Maven, Python, and 30+ other formats. - **Supply chain resilience:** Automated provenance tracking and signature verification help keep you compliant with key security standards and ensure that what you ship is exactly what you built. ## Conclusion: Don’t just move to the cloud – modernize what matters Cloud migration is a rare opportunity to fix the "foundation" of your house before you move in the furniture. By re-evaluating your **artifact management** now, you ensure your cloud-native future is fast, secure, and, most importantly, manageable. The most successful cloud migrations don’t just replicate the past. They modernize the platform that powers everything built next. **Planning a cloud migration?** [Book a demo](https://cloudsmith.com/book-a-migration-consultation) with our experts to simplify the process. ## Frequently asked questions - ### What is artifact management in DevOps? It is the practice of storing, securing, and distributing build outputs, such as Docker images or Maven packages, throughout the development lifecycle. It ensures your builds are reproducible and secure. - ### Why reconsider artifact repositories during cloud migration? Updating your repository during a move avoids duplicate work and ensures that a legacy on-premises artifact manager doesn’t constrain your new cloud infrastructure. - ### What are the risks of keeping a self-hosted repository in the cloud? [Self-hosting](https://cloudsmith.com/blog/cloud-native-vs-on-premise-artifact-management-a-complete-overview) in the cloud still requires manual patching and scaling. This increases costs and creates "visibility gaps" that can lead to security breaches. - ### How does a fully managed artifact repository improve security? Fully managed platforms provide centralized governance, immutable storage, and automated compliance auditing, all of which are critical to a secure **software supply chain**. For years, the bottleneck in software was “how fast can we write code?” Today, Generative AI shifts that bottleneck to **“how fast can we secure it”**. For years, the bottleneck in software was “how fast can we write code?” Today, Generative AI shifts that bottleneck to “[how fast can we secure it](https://cloudsmith.medium.com/is-ai-quietly-making-your-software-supply-chain-less-secure-e1364de33f9a)?” As organizations move from experimentation to production-grade AI, they are discovering that traditional DevOps tooling wasn’t built for a non-deterministic world. For example, static software composition analysis (SCA) scanners that assume deterministic dependency graphs or CI policy gates that validate known build artifacts. When a model generates code rather than a human, the software supply chain changes overnight. Our guide, **Securing non-deterministic systems: A practical guide for AI artifacts and LLMOps**, explores three emerging security frontiers that every organization adopting AI must address: ## 1. AI-generated code introduces supply-chain hallucinations LLMs generate dependencies probabilistically, not deterministically. This creates the emerging **slopsquatting** attack vector, where attackers register hallucinated package names suggested by AI tools and weaponize them with malicious payloads.Without validation and artifact governance, a single copied command can silently compromise an enterprise environment. ## 2. AI models behave like executable software, not passive data Modern model formats can execute arbitrary code during deserialization, most notably through Python pickle-based loading.This **logic-weight entanglement** means downloading an unverified model from public registries such as Hugging Face or Ollama can result in full system compromise.Secure AI development requires scanning, signing, and favoring restricted formats like **safetensors**, alongside enforcing trusted provenance for every model artifact. ## 3. AI productivity and orchestration layers expand the attack surface Frameworks that connect models to enterprise data and automate workflows introduce a new class of high-impact vulnerabilities.Recent RCE exploits in orchestration tools demonstrate that **LLMOps infrastructure itself is now part of the software supply chain**, and must be sandboxed, authenticated, and governed like any production system. ### Ready to harden your AI supply chain? Our full guide provides a strategic roadmap for navigating the shift from DevOps to LLMOps, deconstructing threats in frameworks like Langflow, and building a “sandbox-by-default” development lifecycle. **[Download the full guide: [Securing non-deterministic systems](https://cloudsmith.com/campaigns/securing-non-deterministic-systems-a-practical-guide-for-ai-artifacts-and-llmops)]** **Catch up on the series:** - [_Why Repository Structure Matters?_](https://cloudsmith.com/blog/why-repository-structure-matters) - [_The Hybrid Repository Structure: Balancing Control and Flexibility_](https://cloudsmith.com/blog/the-hybrid-repository-structure-balancing-control-and-flexibility) And now, let’s dive into part three: How **access control and permissions** keep your multi-format repositories secure, consistent, and developer-friendly. In the first two blogs of this series, we explored why repository structure matters and how Cloudsmith’s Hybrid Repository Structure balances control with flexibility. While we touched on policies and permissions, we didn’t dive into the _real_ mechanics of how access control ensures artifact security, traceability, and consistency, especially in **multi-format repositories**, where different packages, languages, and tooling coexist in a single place. While multi-format repositories allow for more flexibility in how your repositories are set up, they can introduce a new way of thinking about how and when different artifacts can be accessed and by whom. This blog breaks down how Cloudsmith provides fine-grained, flexible, and secure controls for teams of any size. ## User roles in Cloudsmith Cloudsmith provides the following user roles: - Owner - Manager - Member - Collaborator These roles help limit access based on organizational need and provide the foundation for more granular permissions. ## Privileges in Cloudsmith - Administrator - Read - Write You can explore the full breakdown of Cloudsmith roles, permissions, and privileges in our documentation. You can read more about user roles, permissions, and privileges in Cloudsmith [here](https://help.cloudsmith.io/docs/access-controls). ## Global privileges in Cloudsmith Every Cloudsmith customer is given the opportunity to set default global privileges. These global default privileges are set for the “Member” user role in Cloudsmith, which is often suited best for individual developers. Within a customer’s global workspace privileges, organizations can choose to grant members the ability to create new teams, invite new users, and even create new repositories. Organization owners can also grant “blanket” repository privileges to all users within Cloudsmith.  While we offer the flexibility for organizations to shift responsibility and access to the developer, we often see our enterprise customers lean away from blanket permissions and access toward more fine-grained permissions. As an example, we have a semiconductor manufacturing customer that has disabled default workspace global privileges so that only Organisation owners are the only users who can invite new users, create new teams, invite 3rd-party collaborators, and create new repositories.  In addition, this customer has access control, and privileges are scoped down to specific teams. Default global repository privileges are disabled so that they can choose exactly which team(s) should have access to repository(s). Even Cloudsmith service accounts are grouped together within a team for ease of tracking permissions and access when it comes to build and deployment times.  ## Repository privileges in Cloudsmith If we zoom in once and look at the repository level in Cloudsmith, we can assign a default privilege for organization members for accessing packages within the repository, and we can assign specific privileges for specific teams, users, or service accounts. This is the stage where you’ll have to consider what packages the repository is storing and if you want developers and service accounts to have default admin, read, or write permissions to the repository. Continuing to use my customer example, they have chosen to set default read permissions for all repositories; however, specific service accounts have write and [admin permissions](https://help.cloudsmith.io/docs/manage-a-repository) to different repositories. You may be okay with your developers downloading artifacts to their local machine for testing or even service accounts requesting stored artifacts tied to staging and production environments. On the other hand, there are customers that may not want their developers having the ability to write to every repository and would most likely want specific service accounts tied to their CI/CD pipelines to only have write permissions. While this is generally best practice, there are certainly exceptions to this rule, as we also have platform teams that choose to grant specific developer teams write access to specific artifact repositories. ## Fine-grained repository controls On top of the permission and access control settings we’ve discussed, Cloudsmith goes even further to ensure that, through various additional settings, platform teams can decide exactly what their developers need and don’t need to be reconfigured within a repository. Platform teams are able to grant or deny developer permissions, such as: - Copying Packages From One Repository to Another - Moving Packages From One Repository to Another - Deleting Packages - Scanning Packages - Replacing Packages - Managing, Using, Or Viewing Entitlement Tokens If you thought that wasn’t enough, Cloudsmith takes it a step further by scoping down permissions to the individual developer’s generated packages. So while platform teams can restrict tampering of artifacts generated by other developers or systems, they can also decide if they would like to allow developers the ability to scan, move, copy, delete, or resync their own packages. Most of the enterprise customers I work with allow developers to do as they please with their own packages to avoid a developer mutiny! In either case, these user actions are catalogued in our [Audit Logs](https://docs.cloudsmith.com/logs-and-observability/audit-logs) for enhanced observability across your Cloudsmith environment. ## Entitlement tokens For instances where our customers want to be very particular about what, how, and when users can access specific packages, Cloudsmith offers [Entitlement Tokens](https://help.cloudsmith.io/docs/entitlements). These scoped tokens are read-access only, so there is never a risk of a user performing a write action against the repository they have limited access to. Customers are able to restrict access by creating a precise search query to narrow down specific packages within a repository, should they choose not to provide visibility into all the packages within the repository. On top of visibility restrictions, customers can also add token usage restrictions to avoid prolonged access and usage of the token. Parameters include: - Token Validity & Expiry Dates - Maximum Downloads - Maximum Clients/IPs - Maximum Download Bandwidth Typically, we see our customers use entitlements for 3rd party software distribution or even for systems that don’t require logins to Cloudsmith. ## Geo/IP Restrictions For our security-conscious customers, Cloudsmith also offers the ability to configure [geo-based restrictions](https://help.cloudsmith.io/docs/geoip-restriction) based on country, with an easy-to-use preconfigured list of countries to choose from. Choose to deny or allow access from specific countries. Although keep in mind that theoretically, no unauthenticated user should have access to your Cloudsmith workspace in general. This restriction applies more towards open-source repositories that you may be hosting in Cloudsmith, but it’s always a good idea to practice defence-in-depth! If geo-based restrictions are too broad, you can scope down to IP-based restrictions to either allow or deny client access based on IP address. This added protection ensures that requests coming from clients with an unapproved IP address do not have access to your repositories.  ## We’re here to help With all of these configuration options, it can be tricky to decide how you want to best enable your developers while balancing security and flexibility. Not to worry—the Cloudsmith Customer Success team is here to help you in your decisions throughout the onboarding process. We’ve walked through these decisions with many customers and have outlined decision impacts for you so you’re not left wondering “what if”. Learn more about Cloudsmith [here](https://cloudsmith.com/) to get started with us. See you on the other side!  ## FAQs (Frequently asked questions) ### 1. What is access control in a multi-format repository? Access control defines who can read, write, or administer artifacts across different package formats stored in the same repository. It ensures secure, consistent governance across diverse tooling. ### 2. How do permissions work in Cloudsmith for multi-format repositories? Cloudsmith supports global, repository-level, team-based, and user-specific permissions, allowing organizations to tailor access to packages, teams, service accounts, and even individual artifact actions. ### 3. Why is fine-grained access control important in software supply chain security? Fine-grained controls reduce the blast radius of errors, prevent unauthorized writes, and help organizations enforce policies required by modern software supply chain frameworks like SLSA and SSDF. ### 4. What are entitlement tokens used for in Cloudsmith? Entitlement tokens provide scoped, read-only access to specific artifacts without requiring user accounts, making them ideal for external distribution, automation, or least-privilege workflows. ### 5. Can developers manage their own packages in Cloudsmith? Yes. Cloudsmith allows permissions that let developers manage only the packages they personally created—without putting others’ artifacts at risk. ### 6. How does Cloudsmith support secure CI/CD pipeline access? Service accounts can be granted precisely the permissions needed for pipeline operations (like write or admin) while keeping developers and collaborators restricted to read-only or scoped access. ### 7. What’s the difference between global and repository-level privileges? Global privileges affect the entire workspace, while repository-level privileges enable granular control over specific teams, users, or service accounts interacting with a particular repository. In a world where software ships in seconds, teams are still chained to legacy systems built for a different era. What once passed as “good enough” for storing and distributing builds has become a drain on productivity - adding risk, slowing delivery, and quietly inflating costs year after year. In this post, we’ll break down the hidden cost of legacy artifact repositories, discuss the importance of modernizing through cloud-native artifact management, and demonstrate how you can leave the old infrastructure that has been slowing your software supply chain. ## What is legacy artifact management? Legacy artifact management involves older on-premise artifact repositories or in-house custom systems. These tools were designed in another era, when teams used monolithic applications and updates were done once a year or even less. The modern reality is very different. Cloud-native development, continuous CI/CD pipelines, and distributed engineering teams need a modern artifact management approach that delivers scalability, uptime, and built-in security. This is where legacy artifact repositories fall behind. Many teams assume that on-prem systems are “more secure” because they’re isolated, but isolation no longer protects against today’s threats. Most attacks now originate upstream, through open-source dependencies that already contain vulnerabilities, malicious code injections, or compromised packages.  With the volume and speed of issues emerging in the open-source ecosystem, an isolated, self-hosted repository cannot keep pace without continuous scanning, real-time visibility, and automated updates. Without these protections, legacy artifact management becomes a blind spot in the software supply chain - quietly storing and distributing unverified or unsafe artifacts. ## The hidden costs of legacy or on-premise artifact repositories Legacy systems can feel safer to stick with - they’re already in place and “working.” But maintaining the status quo often hides bigger costs: outdated infrastructure, ongoing maintenance, and mounting security risks from unpatched or unsupported components. ### 1. Complexity and maintenance overhead Legacy repositories need to be manually patched, updated, backed up, and scaled. Teams spend valuable engineering time on server management instead of innovation. Each new project or environment increases the complexity of configuration and slows down the development. ### 2. Unrecognized infrastructure costs Hosting artifact repositories either in on-prem or in self-managed cloud VMs requires continued expenditure on storage, bandwidth, and compute. The costs increase unintentionally as the size and volume of artifacts grows (especially large Docker images or build artifacts). Beyond infrastructure, many older systems also require costly vendor support contracts for upgrades, patches, and troubleshooting. These fees often increase over time and are non-negotiable. ### 3. Security and compliance risks Legacy systems often lack built-in scanning, access control, or software bill of materials (SBOM) capabilities. Lacking transparency in dependencies leaves teams exposed to vulnerabilities and unmet compliance requirements - something that enterprises will not be able to afford in 2025 and beyond's regulatory environment. ### 4. Scalability limitations As repositories grow, performance bottlenecks emerge - developers face slower downloads, failed builds, and pipeline downtime, all of which directly slow release velocity and drain productivity. ## Why cloud-native artifact management is a change worth making Legacy tools may have provided the groundwork for early DevOps, but they cannot keep up with the current software landscape. Modern organizations need next-generation artifact management - cloud-native solutions designed for speed, security, scalability, and seamless integration with cloud-native CI/CD pipelines. A [cloud-native artifact management](https://cloudsmith.com/product/cloud-native-artifact-management) platform is more than just a storage location for packages - it’s a critical pillar of your software supply chain security. It guarantees that every artifact, from source to deployment, is verified, traceable, and instantly accessible, regardless of where your teams are working. These capabilities are exactly why more organizations are moving to modern, cloud-native platforms that combine speed, security, and scalability to support today’s software delivery demands - reasons we explore in detail below. ### 1. No maintenance and always up to date With legacy artifact repositories, engineering teams spend hours managing servers, applying patches, and juggling storage. A truly [cloud-native artifact management platform](https://cloudsmith.com/blog/artifact-management-a-complete-guide) is different from simply hosting a repository in the cloud—it’s built to auto-scale, self-update, and deliver continuous security without manual intervention. There’s no server downtime, no upgrade windows to schedule, and no need to plan for storage expansion - everything is handled seamlessly in the background. ### 2. Scalability without complexity Self-hosted systems cannot keep pace with the growth in artifact volume size. A cloud-native artifact management platform dynamically boosts its capacity to manage millions of artifacts across multiple teams, regions, and projects, [without compromising performance](https://cloudsmith.com/product/global-software-distribution). Using elastic storage and edge caching CDNs, developers are always guaranteed a [quick download and high uptime](https://cloudsmith.com/blog/scaling-for-extreme-performance). ### 3. Built-in security and compliance from the ground up In today’s world, threats move faster than ever. With malicious packages and supply chain attacks on the rise, cloud-native artifact repositories integrate vulnerability scanning, access controls, and SBOMs (Software Bill of Materials) directly into your pipelines. This ensures that all artifacts maintained and shared are validated, trackable, and consistent with the industry regulations such as SOC 2, ISO 27001, and FedRAMP. No extra patching or standalone security tools - modern artifact management [integrates security](https://cloudsmith.com/product/software-supply-chain-security) into every phase of your software supply chain. ### 4. Performance and speed that empower developers A truly cloud-native artifact repository ensures artifacts are forwarded to the closest edge location, which significantly decreases both the time spent building and deploying, which has a direct impact on increasing developer productivity and CI/CD throughput. Engineers can focus on building features rather than waiting on downloads or troubleshooting failed builds, making software delivery faster, more reliable, and predictable. ### 5. Seamless integration with the modern DevOps toolchain Legacy repositories often require plugins or manual scripting to integrate with CI/CD tools. Cloud-native artifact management platforms offer native integrations with [GitHub Actions](https://cloudsmith.com/product/integrations/github-actions), [GitLab CI](https://cloudsmith.com/product/integrations/gitlab-cicd), [Jenkins](https://cloudsmith.com/product/integrations/jenkins), [CircleCI](https://cloudsmith.com/product/integrations/circle-ci), and more - all via robust APIs. This ensures that artifacts flow seamlessly through your CI/CD pipelines, maintaining consistency, traceability, and reliability from development all the way to production. ### 6. Single visibility and governance between teams In large organizations, artifacts are often scattered across multiple repositories and sometimes duplicated, making governance and visibility a constant challenge. A cloud-native artifact management system provides a centralized, single platform for managing visibility, audit, and access. Administrators can also manage published, promoted, or consumed artifacts - to ensure compliance and reduce the risk of unauthorized access or obsolete dependencies. ### 7. Predictable, transparent costs In contrast to self-managed solutions that have unpredictable infrastructure charges, cloud-native artifact management follows a usage-based pricing scheme, which is predictable. You pay for what you use. You do not expect to incur costs for hardware, maintenance, or downtime. This will ultimately lead to a reduced total cost of ownership (TCO) and a better understanding of the ROI of engineering time. Moving legacy artifact management to a new, modern, cloud-native repository is not just a technical choice but a strategic one that enhances both the security, performance, and user experience of developing a product or service, as well as reduces costs in the long run. Migrating from legacy artifact management to a modern, cloud-native repository improves security, performance, and the overall developer experience while helping reduce long-term operational costs. By centralizing control, simplifying scalability, and strengthening the software supply chain, teams can focus on building software more efficiently and securely. ## How to upgrade from escape legacy artifact management (step-by-step) The idea of moving away from on-premise or legacy artifact systems to a modern, cloud-native solution can be overwhelming, but with the correct plan, it is achievable. 1. **Audit your existing repositories – **Review what you store (packages, containers, Helm charts, etc.) and where. 2. **Analyze utilization and access patterns -** Learn which teams, pipelines, and tools rely on which repositories. 3. **Select a modern artifact management platform – **Seek capabilities such as universal format support, security scanning, policy management, global availability, and automation through integrations. 4. **Plan your migration strategy – **Migrate critical projects first, automate uploads, and validate integrations. 5. **Decommission legacy infrastructure – **Once migration and validation are complete, phase out outdated systems to eliminate ongoing maintenance, reduce operational overhead, and free up resources for modern, cloud-native artifact management. 🔥**Top tip: **Cloudsmith’s [Migration](https://cloudsmith.com/campaigns/cloudsmith-migration-guide) Toolkit, combined with expert support, makes the transition seamless - preserving your history and metadata while enabling improved security and scalable infrastructure. ## The real ROI of leaving legacy, on-premise artifact management behind Teams that modernize and migrate to cloud-native artifact management see measurable returns: - Reduce infrastructure expenses by up to 60%. - Faster build and deployment times across CI/CD pipelines. - Improved developer satisfaction through simplified workflows. - Better compliance posture with automated vulnerability management. Time spent maintaining a legacy repository directly impacts productivity and costs. Migrating to a modern, cloud-native artifact repository preserves operational efficiency and supports long-term software delivery improvements. ## How Cloudsmith makes modern artifact management effortless All of these challenges, including scalability, security, automation, and visibility, can be solved with a truly cloud-native approach to artifact management. And if you are planning a migration to the cloud, it is worth doing it right rather than sticking with your existing provider simply because it feels easier. A migration is already a major change, and it is the perfect opportunity to elevate your entire artifact management program. Cloudsmith was built from the ground up as a fully cloud-native platform that helps teams break free from the limits of traditional repositories, delivering seamless automation, built-in security, and scalable reliability in a single unified system. Here’s how Cloudsmith enables that transition seamlessly: - **Fully managed, always available:** Cloudsmith is truly cloud-native which means hosting, scaling, and security are built in. Teams can focus on development without worrying about infrastructure maintenance and downtime. - **Universal support for all formats:** Whether you manage containers, packages, Helm charts, or custom binaries - with multi-format repositories, Cloudsmith provides one centralized platform for all your artifacts. - **Unified security and compliance:** Each artifact is scanned, signed, and tracked. Cloudsmith also has vulnerability scanning, dependency metadata, and SBOM generation built-in to ensure end-to-end security for your software supply chain. - **Global performance and distribution:** Artifacts are served over Cloudsmith’s global edge network and minimizing latency and providing fast and reliable builds across the globe. - **Seamless CI/CD integration:** Cloudsmith integrates seamlessly with the latest DevOps platforms: GitHub Actions, GitLab CI, and Jenkins - enabling teams to automate artifact workflows, reduce manual errors, and accelerate software delivery. A modern, cloud-native artifact repository like Cloudsmith simplifies operations, strengthens security, and accelerates software delivery - without the hidden costs or complexity of legacy systems. ## Summary: don’t let legacy on-premise artifact management hold you back Legacy artifact management is not only dated - it’s also costly, risky, and non-sustainable. The emerging generation of cloud-native artifact management platforms, such as Cloudsmith, transcends complexity with confidence, enabling teams to achieve the visibility and velocity required to build securely at scale. The faster you retire legacy systems, the sooner your organization can build securely, on a truly modern, cloud-native platform. ## Frequently asked questions (FAQs) #### 1. What is legacy artifact management, and why is it a problem? The management of legacy artifacts encompasses older systems (typically [on-premise artifact management](https://cloudsmith.com/blog/cloud-native-vs-on-premise-artifact-management-a-complete-overview)) used to store and distribute software packages. Such systems do not offer the automation, scalability, and integrated security that are needed in modern DevOps, resulting in inefficiencies and increased operational costs. #### 2. What are the unknown expenses of legacy artifact repositories? Beyond licensing, teams also bear the costs of infrastructure maintenance, downtime, manual updates, and security risks. These hidden expenses can quickly add up, often exceeding the investment required for a modern, cloud-native alternative. #### 3. How do I migrate to a modern artifact repository? Begin by auditing your existing repositories, determining dependencies, and automating the migration with the help of migration tools (such as Cloudsmith’s Migration Tool). The advantage is to retain the integrity of artifacts with the purpose of avoiding manual management. #### 4. Why choose a cloud-native artifact repository over self-hosted options? Uptime, scaling, and security are automatically managed on cloud-native platforms. They are CI/CD integrated, can distribute faster worldwide, and can eliminate maintenance overhead, allowing your developers to focus on their core tasks #### 5. How does modern artifact management improve security and compliance? Modern artifact systems integrate vulnerability scanning, SBOMs generation, and access control. This will ensure artifacts are secure, traceable, and compliant – which is essential to securing your software supply chain. At Cloudsmith, using Multi-format repositories, we provide a simple and flexible solution to deploy and distribute your software artifacts. Multi-format repositories allow you to store artifacts of different formats in the same place. Organize your packages by environment, project, package type, or whatever way you see fit- we are not opinionated about how you organize your packages or containers. Cloudsmith is the most flexible, universal package manager on the market, with Multi-format repositories and support for all package formats. ## **Diverse Tech Stacks** In the not-too-distant past, it was common to see organizations characterize themselves as Java or C# houses- it’s rare nowadays to find a project that is developed entirely in a single language. A common project may use React for server-side rendering, Kubernetes, and Docker for orchestrating microservices and have microservices built using Python, Java, and Node.js. Different frameworks and languages are used depending on the expertise and problem they are tackling- backend, frontend, containerization- there are loads of good reasons to have multiple languages in your project- Who are we to judge your tech stack! Many package managers don’t support a broad range of packages. That means you’ll require several repositories (Public or Private) and therefore several places that you need to manage, several places that you’ll need to integrate with your build processes, and several places that you’ll have to ensure your team has reliable, performant access to. Other package managers don’t let you store different packages in the same repository. They make you create a repository for all your NPM packages, another for your Maven packages, and another for your Docker images. At Cloudsmith, you can do this no problem, but you don’t have to 😊 ## **Multi-format repositories** What exactly is a Multi-format Repository? Cloudsmith’s Multi-format repositories can house all your packages in one single location- no matter what package format you are dealing with! You can throw your Docker images in with your NPM packages, your NuGets with your Helm Charts- if it makes sense for your project and your team, go for it. Your Multi-format repository will still work as expected with your native packages management tooling. You can use the native tooling for all supported packages- ‘docker pull’, ‘pip install’, or ‘npm install’ for example- all from the one single repository! So while you have a single repository and a single set of processes for managing, sharing, and controlling your software assets, you lose absolutely nothing when it comes to functionality. The benefits of Multi-format repositories aren’t just limited to development environments either. If you are distributing your software to your own customers, and you provide it in a number of different formats, wouldn’t it be nice to offer all those formats to your customers from a single location? You no longer have to worry about pushing/publishing to multiple different repositories. Multi-format repositories make it easy to develop, deploy, and distribute software. ## **Universal Package Tagging** If you are worried about losing some information by changing your repository structure, don’t worry- you can enrich your packages with Universal Package Tagging. Cloudsmith Multi-format repositories mean you can apply these tags across all the package formats you use, all in one repository. With universal package tagging, you now have the ability to add your own searchable attributes to your packages without complicating the repository structure. ## **Example repository structures, please** We are not opinionated around how you organize your repositories. There are many ways to organize your repositories- below, we explore how you can structure your repositories around logical groupings of environments, projects, or package type. ### **Environmental specific repositories** A popular structure for customers in Cloudsmith is to have your repositories mirror your production pipeline and then use webhooks to automate deployment and promote assets. A typical setup is to have a development, test, staging, and production repository. Once you upload an asset, you can promote it through the repositories: development -> test -> staging -> production. This structure helps keep the integrity of the asset, a provenance trail, and without the bandwidth cost of downloading from staging then re-uploading to production. Nice and simple, easy to automate against- it’s a classic! ### **Project-specific repositories** I love seeing all the packages associated with a project in the same place, especially when working on a project which uses a microservices architecture with a diverse tech stack. Use your repository to help Developers visualize how the project works together. Storing all your packages associated with a project in the same repository is a simple and practical way to organize your packages. ### **Package specific repositories** Of course, if you want to create a repository for each format you use, you absolutely can do that - there is nothing about Cloudsmith’s Multi-format repositories that would prevent you from doing so. We like this structure, and it fits beautifully into many DevOps processes, but we don’t like that other providers limit you only to use this structure. ## **Multi-format is the best fit for modern tech stacks** Modern tech stacks use multiple languages, frameworks, and tools. Cloudsmith’s Multi-format repositories are ideally suited to this environment- a flexible universal package manager that can store any package format in whatever repository structure you choose fit. Reduce the complexity of deploying software and gain the flexibility to structure your repositories to suit your team and processes using Multi-format repositories. ## Introducing Terraform [Terraform](https://www.terraform.io/) is an awesome tool, created by Hashicorp, with a vibrant community of open source modules for multiple cloud providers. It’s not hard to understand why it’s a favourite amongst developers; especially when it's often faster at supporting new features than some of the cloud providers! Terraform enables its users to manage components in their architecture, ranging from the configuration of the services used in their chosen cloud provider (such as [Amazon Web Services, Google Cloud Platform, or Microsoft Azure](https://www.terraform.io/docs/providers/type/major-index.html)), to software run on those services (like [Chef, Helm or Vault](https://www.terraform.io/docs/providers/type/infra-index.html)). Terraform even supports configuring some SaaS services you may interact with (for example, [GitHub or GitLab](https://www.terraform.io/docs/providers/type/vcs-index.html)). Using Terraform to define your desired configuration means that not only will you offload the work of creating the components you need, but also make the creation of these instances reproducible, allowing you to create, update and destroy components of your architecture as needed, all with one command (and a review of the changes before executing them!). ## The private module registry To make modules created by the Terraform providers available to all Terraform users, Hashicorp runs the [public Terraform Registry](https://registry.terraform.io/) which allows users to upload modules, making them publicly available to all Terraform users. Cloudsmith's private module registry fully supports the [Registry API](https://www.terraform.io/docs/registry/api.html), allowing users to access both private modules available via Cloudsmith and public modules via the public registry. Cloudsmith's private module registry also enables developers to: - Share custom modules for managing infrastructure among their team - Utilize private modules in your organization without the overhead of managing their storage - Distribute private modules as a commercial offering - Make modifications to public packages while having control of how you publish them (publicly accessible, private to your organization or open-source) - Capture the exact state of modules you depend on at a particular version - Control (allow list/deny list) at an organization, repository, and module level ## Getting Started Getting started with Cloudsmith and Terraform only takes a few steps. First, you need to create a Cloudsmith account and a repository to which you can upload your modules (a terraform provider for doing just this coming soon!). If you need to install Terraform, you can find a getting started guide on the [Terraform website](https://learn.hashicorp.com/terraform/getting-started/install). Cloudsmith's private module registry should work with all supported versions of Terraform, however, we recommend using version **0.12 or higher**. You can check the version you have installed by running: ```json { "_key": "0ed3c0e9234a", "_type": "code", "code": "$ terraform version\n\nTerraform v0.12.26", "filename": null, "language": "text", "markDefs": null } ``` ### Creating a module For this example, we'll create a module which prints out "Hello, world!". First, we need to create the directory which will contain the module: ```json { "_key": "963410dc184c", "_type": "code", "code": "$ mkdir my_module && cd my_module\n", "filename": null, "language": "text", "markDefs": null } ``` Following this, we need to create the `main.tf` file, this is the entry point for your module and will contain the code required to generate our message. To generate this message we must add the following to the `main.tf` file: ```json { "_key": "aeba04884db0", "_type": "code", "code": "resource \"null_resource\" \"message\" {\n provisioner \"local-exec\" {\n command = \"echo 'Hello, world!'\"\n }\n}", "filename": null, "language": "text", "markDefs": null } ``` To see the output, you can run: ```json { "_key": "9aeea8a1e38b", "_type": "code", "code": "$ terraform init && terraform apply", "filename": null, "language": "text", "markDefs": null } ``` ### Uploading your module Now that we have the code necessary for our module, we need to package and upload it to Cloudsmith. To achieve that we need to make use of the *nix tool `tar` to package up our module: ```json { "_key": "1dd9d8417bb6", "_type": "code", "code": "$ tar --exclude='.terraform' \\\n --exclude='.terraformrc' \\\n --exclude='*.tfstate' \\\n --exclude='*_override.tf*' \\\n -zcf \"terraform-local-example-v0.0.1.tar.gz\"", "filename": null, "language": "text", "markDefs": null } ``` To ensure we don't upload some unnecessary files, we exclude your terraform state file(s) along with `.terraformrc` which can contain authentication credentials. The name of the terraform module is critical here, as the [terraform registry API requires this specific format](https://www.terraform.io/docs/registry/modules/publish.html#requirements), along with several other requirements that you should review when creating modules. To upload the module to Cloudsmith, you can use either the application's UI, or the Cloudsmith CLI. For this example, we'll be using the CLI, which has steps for installation available [here](https://help.cloudsmith.io/docs/cli). With the CLI installed and configured to authenticate successfully with Cloudsmith, we can upload our module by running: ```json { "_key": "b1234df5ef39", "_type": "code", "code": "$ cloudsmith upload terraform your-namespace/your-repository terraform-local-example-v0.0.1.tar.gz", "filename": null, "language": "text", "markDefs": null } ``` Once uploaded, you can view your module in Cloudsmith. ```json { "_key": "4a304f43cb5a", "_type": "image", "alt": null, "asset": { "_createdAt": "2025-06-05T07:58:55Z", "_id": "image-9a61233e3760189e129970cbbb76ee73a0c12930-1166x955-png", "_rev": "gnJeqIUmT5gWK5E6lfoHB7", "_type": "sanity.imageAsset", "_updatedAt": "2025-06-05T07:58:55Z", "assetId": "9a61233e3760189e129970cbbb76ee73a0c12930", "extension": "png", "metadata": { "_type": "sanity.imageMetadata", "blurHash": "e4SF@UICJA~XNJ0#rx9ZRPNKkq%3M_xZt8_2IU-;tQV[=xX3~Vt6V?", "dimensions": { "_type": "sanity.imageDimensions", "aspectRatio": 1.2209424083769633, "height": 955, "width": 1166 }, "hasAlpha": true, "isOpaque": true, "lqip": "data:image/png;base64,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", "palette": { "_type": "sanity.imagePalette", "darkMuted": { "_type": "sanity.imagePaletteSwatch", "background": "#544e5f", "foreground": "#fff", "population": 0.02, "title": "#fff" }, "darkVibrant": { "_type": "sanity.imagePaletteSwatch", "background": "#1e4666", "foreground": "#fff", "population": 0, "title": "#fff" }, "dominant": { "_type": "sanity.imagePaletteSwatch", "background": "#367fb8", "foreground": "#fff", "population": 0.33, "title": "#fff" }, "lightMuted": { "_type": "sanity.imagePaletteSwatch", "background": "#a2b6c9", "foreground": "#000", "population": 0.29, "title": "#fff" }, "lightVibrant": { "_type": "sanity.imagePaletteSwatch", "background": "#ecbe84", "foreground": "#000", "population": 0.12, "title": "#fff" }, "muted": { "_type": "sanity.imagePaletteSwatch", "background": "#a75c50", "foreground": "#fff", "population": 0.01, "title": "#fff" }, "vibrant": { "_type": "sanity.imagePaletteSwatch", "background": "#367fb8", "foreground": "#fff", "population": 0.33, "title": "#fff" } } }, "mimeType": "image/png", "originalFilename": "Screenshot-2020-06-18-at-16.19.09.png", "path": "images/rafvlnhi/production/9a61233e3760189e129970cbbb76ee73a0c12930-1166x955.png", "sha1hash": "9a61233e3760189e129970cbbb76ee73a0c12930", "size": 144710, "uploadId": "CN367amFEL8IKfZdhuV3rzvoducVBrCn", "url": "https://cdn.sanity.io/images/rafvlnhi/production/9a61233e3760189e129970cbbb76ee73a0c12930-1166x955.png" }, "caption": null, "link": null, "markDefs": null } ``` ### ### In summary Cloudsmith provides fully featured Terraform module repositories on all plans, flexible enough for use whether you’re hosting public packages for a public or open-source project, or private packages for your company’s internal needs. We're extremely proud to be able to support the Terraform ecosystem with this tooling. You can find further, context-specific information, including detailed setup and integration instructions inside each Cloudsmith repository. Why wait? Get your public and private [Terraform package repository](/product/formats/terraform-registry) hosting at Cloudsmith now. The movement away from on-premise and towards the Cloud is unstoppable. Even the [US](https://www.whitehouse.gov/briefing-room/presidential-actions/2021/05/12/executive-order-on-improving-the-nations-cybersecurity/) government is on board with their plans to “accelerate movement to secure cloud services, including Software as a Service (SaaS), Infrastructure as a Service (IaaS), and Platform as a Service (PaaS).” There's many types of infrastructure to consider when selecting software. We aim to uncover the true difference of the key two, being on-premise software and cloud-native software below. ## What is cloud-native software? Cloud-native software and technologies are an approach to architecting, building, and managing workloads built in the cloud. This means that cloud-native software has the Cloud inherently built into its DNA, taking full advantage of all the technology that Cloud provider platforms offer. This results in the ability to scale for global performance. ## What is on-premise software? On-premises (or on-prem) software uses IT infrastructure hardware and software applications that are hosted on-site, and relies on constant maintenance and upkeep from dedicated resources. On-prem is quickly becoming the legacy approach in artifct, as organizations are realizing the power of the Cloud. ## **What exactly is the difference between on-premise and cloud-native?** ### On-Premise On-premise (on-prem) software is deployed, hosted, and maintained by your organization. While you can host your on-prem software in the cloud to remove some infrastructure and maintenance costs, you really need cloud-native software to take full advantage of the Cloud. ### Cloud-native Cloud-native software is designed, developed, and deployed to take advantage of modern cloud computing, using techniques like micro-services, modularity, containerization, continuous integration, and continuous deployment. When you access your software over the cloud, that usually means using the SaaS licensing and delivery model. You access your software over a browser or via an API, and you pay for what you use, like a utility bill. Cloud-native software using SaaS licensing empowers you to focus on growing your business while only paying for the services you use. ## Advantages of Cloud-Native Software over On-Premise Software Many new companies today are choosing to offload the overhead of maintaining the tools and infrastructure they rely on to cloud-native services. The use of on-prem software is mainly due to legacy or regulatory reasons. For example, the innovation coming out of Cloud-first FinTech companies compared to the risk-averse old-school banking industry shows what you can do when unshackled from your on-prem mindset. Now you see older banks like [Morgan Stanley](https://www.forbes.com/sites/tomgroenfeldt/2021/06/03/morgan-stanley-and-microsoft-partner-for-technology-modernization/?sh=2871942d197b) accelerating their use of the cloud to remain competitive. There are many advantages of cloud-native over on-premise software: ### Cost of Maintenance and Infrastructure Cloud software is hosted for you. You don’t have to worry about maintaining your “on-prem” software or infrastructure- No updates, no security patches, no replacing obsolete hardware. Cloud-native technologies allow businesses to reduce the total cost of ownership for businesses, especially when you factor in the staff costs of maintaining “on-prem,” never mind just the licensing fees. ### Scalability Scalability is another area where the cloud excels. Cloud-native software can quickly re-adjust its resources to meet demand. A company experiencing rapid growth can use the cloud to expand its infrastructure and computing power. In contrast, the same company using on-prem infrastructure would have to quickly invest in more hardware, software, and Engineers to keep up with rapid growth. On-prem software cannot compete with Cloud-native software in terms of scalability and flexibility. ### Distributed Teams “On-prem” software solutions tend to be faster the closer you are to the infrastructure. It’s no longer acceptable for everyone near the company headquarters to have a responsive low latency experience, while other geographically distributed teams have to put up with significant replication lag. This lag stifles collaboration and productivity. Cloud software can use techniques like content delivery networks (CDN) and edge caching to provide better performance for distributed teams without requiring you to implement complex global replication or maintain globally distributed infrastructure. ### Security A secure system needs a secure building, training, constant security updates, high availability, monitoring, and disaster recovery infrastructure. Although many companies that host their software on-prem take security very seriously, it is expensive and consumes many working hours. Cloud providers are driven to focus on security as their business and reputation depend on providing a robust and secure service. As a result, cloud providers use highly sophisticated security tools and resources beyond the reach of most in-house teams. ### Where are the Software Devs? The [US Bureau of Labor Statistics](https://www.bls.gov/ooh/computer-and-information-technology/software-developers.htm) projects that from 2019 to 2029, there will be a 22% increase in the number of jobs available to Software Developers. It’s already hard to find Software Engineers, and the best ones want to work with the latest technologies. We need to embrace tooling that takes advantage of the automation and scalability of cloud-native technologies- freeing up your engineering and server resources to build your products. ## Cloud-Native, not just Cloud-Hosted When moving from “on-prem,” it’s crucial to select cloud-native tools, like [Cloudsmith](https://cloudsmith.com/) to use with your cloud infrastructure. Some vendors offer their on-prem software on Cloud platforms. However, these legacy offerings were not designed to run on a Cloud platform or architected to take advantage of the paradigm shift that cloud computing enables. You are left with effectively paying on the double for the same on-prem software, just running in a virtual machine in the Cloud. These Frankenstein hybrid software solutions inherit the same scale and distribution problems as their on-prem alternatives. Cloud-native software works in tandem with Cloud Infrastructure- secure, fully distributed workloads, fault tolerance, and thanks to containers, it can move from cloud to cloud with ease and scale in or out rapidly. ### Cloudsmith- your very own Cloud-Native packaging tool Package management is an example of a tool that would benefit from being Cloud Native. Package management tools develop, deploy and distribute software packages. Cloudsmith is a Cloud-native package management tool that makes life simpler for your engineers. Don’t worry about infrastructure, patching, upgrades, replications, or scaling. Our cloud-native architecture enabled us to develop a smart CDN for software packages- or our Package Delivery Network (PDN). The PDN is optimized to ensure lightning-fast delivery for deploying or shipping licensed software to your customers. We believe that the elasticity, flexibility, security, and scalability of the Cloud allow us to support infrastructure and service beyond the capability of traditionally provided “on-prem” tools. Shift to the cloud away from on-prem software by choosing Cloud-native software to empower innovation, stay secure, reduce costs, and scale as your business needs change. In early 2020, threat actors breached the build systems of SolarWinds and used this access to add malicious code into one of SolarWinds products. The product, called “Orion”, is very widely used and deployed by tens of thousands of companies, including many Fortune 500 companies. Then, from around March 2020, when SolarWinds’ customers updated their installations of Orion, the malicious code embedded gave the threat actors a backdoor which they could then use to install additional malware, spy on, and exfiltrate data. By now, this is a very well-known, maybe even the most famous, example of a software supply chain attack. ## **What has happened since then?** In short – a lot! Unfortunately, the SolarWinds’ incident was not the last time a software supply chain was breached. Software supply chain attacks have been continuing and could be accelerating. In February, it was discovered by security researchers that many public package repositories and registries could be used to implement a [Dependency Confusion attack](/blog/dependency-confusion-attacks), which allowed the insertion of malicious code or packages into several major companies’ build and development processes, including Apple and Tesla. Then in May, it was announced that [threat actors had breached a very commonly used development tool, CodeCov](/blog/thoughts-on-the-codecov-breach). And just a few days ago, it was revealed that the primary public registry for Winget, Microsoft’s new Windows 10 package manager, was being filled with [corrupt, duplicate, or possibly malicious](https://www.bleepingcomputer.com/news/security/windows-10s-package-manager-flooded-with-duplicate-malformed-apps/) packages. ## **Is this new?** Despite the amount of news that these incidents have generated recently, unfortunately, software supply chain attacks are not new. In fact, in 2015, a [fake version of Apple’s XCode](https://en.wikipedia.org/wiki/XcodeGhost) development environment was distributed, which, when used to develop software for iOS, inserted malicious code into dozens of iPhone apps. And in 2019, the same thing happened to Microsoft’s Visual Studio, which resulted in [malware being inserted](https://www.wired.com/story/supply-chain-hackers-videogames-asus-ccleaner/) into several games developed using it. In 2017, another software supply chain breach launched the [Cyber Attack known as NotPetya](https://www.bloomberg.com/news/articles/2017-06-28/microsoft-analysts-see-hack-origin-at-ukrainian-software-firm). The software updates for a Ukrainian accounting product were compromised and caused an estimated $10 billion of damages. And also, in 2018, a group of threat actors compromised software updates from the manufacturer ASUS and the computer cleanup tool CCleaner. These particular attacks are believed to be [just two of six](https://www.wired.com/story/barium-supply-chain-hackers/) successful supply chain attacks perpetrated by this group. Overall, these attacks have affected thousands, if not hundreds of thousands, of users and systems. And often, the difficult part is determining exactly how large the blast radius is. ## **Why is this happening more now?** There are a few reasons why we see an increase in software supply chain attacks, so let’s just touch on a couple of the most important ones. The first reason is that software security is not new. For years now, we have been conditioning users to “not click links in unsolicited emails”, “Check that the site you are using is genuine and secure”, “protect username/password/credentials”, etc. As users, we have started to implement better opsec; we are more aware of threats and how to combat them. So, as a result, the threat actors have had to move up the chain. Rather than target individual users and systems, they target the very processes by which users obtain their software. After all, who wouldn’t trust a software update that has been delivered from a software vendor's distribution platform and signed by the said vendor? Everything we have been conditioned to question and check passes in this case. The second reason is that the blast radius is potentially so much greater. Rather than targeting thousands of users or systems and compromising just a few, an attack like that on Solarwinds results in a single compromise expanding to many thousands of compromised systems and networks. The threat actors can leverage the distribution networks of a software vendor to do the dirty work for them. To let thousands of Trojan horses out into the wild and run wild. The payoff is just so much greater for a successful supply chain attack, that it represents a significantly larger return on investment to an attacker. ## **What can we do about it?** At Cloudsmith, we believe that the solution comes in the form of Continuous Packaging. Continuous Packaging is a development methodology that offers the observability and control to ensure that software is always verified, packaged, and ready to deliver. We have talked about this before, and you can watch our talks on continuous packaging on [our YouTube channel](https://www.youtube.com/playlist?list=PL6gR0yh_5JnoIPKxfwUba4QrD9MbH1fh6). We think that continuous packaging is the missing link between continuous integration and continuous deployment – The cornerstones of a modern DevOps software development process. It really comes down to using new best-in-class tools to help you have visibility and management over all packages and code used in your environments—providing trust, isolation, and performance. Some key concepts to note: · Minimise trust. Scan at source, during build and at packaging. · Isolate from third parties. · Implement and own a single source of truth for packages. · Be a curator of software packages, not just a consumer. · Draw a thread from build to deployment. We will dig a bit deeper into continuous packaging in an upcoming blog, so check back soon and watch this space. However, we are not alone. The [US and UK governments](/blog/united-in-security-us-and-uk-vow-to-improve-software-supply-chains) have recently announced efforts to improve the security of software supply chains. These efforts should result in guidelines for best practices and minimum standards that vendors should meet. At its core, this is what Cloudsmith does. Cloudsmith provides a single source of truth for all packages and containers you consume as part of your development process. It enables you to isolate yourself from public upstream package sources and take back control over your software supply chain. It’s time for a change. It’s time for an overhaul of how we securely develop software, and it’s time to get ready.