Skip to main content
CleanStart

Knowledge Hub

The Numbers

Enterprise Scale

CleanStart operates at enterprise scale, delivering verified container images across complex software ecosystems spanning multiple programming languages and package managers. These numbers represent real, tracked data that powers the security intelligence and analytics at the core of the platform.

Image Coverage

Metric

Value

Scope

Pre-built images

1,200+

Ready to use out of the box

Language variants

19,200+

Node, Python, Go, Java, Rust, Ruby, C/C++ and more

Architecture support

2

AMD64 + ARM64 (native compilation, not cross-compiled)

Test scenarios

78

Security, functionality, performance tests per image

Total test runs

1,574,400+

78 tests × 19,200+ variants

For any language version and library combination you use, CleanStart has likely pre-built a verified variant. If not, you can declare it via YAML configuration and get a hermetic build in 30-60 minutes.

Dependency Intelligence

The supply chain security model rests on understanding dependencies at scale.

Dependency Graph

Metric

Value

Context

Tracked dependencies

281M+

Libraries across 7 ecosystems

Ecosystems

7

Go (238M), Crates (16.6M), npm (12.5M), Maven (7.2M), PyPI (3.4M), RubyGems (3.1M), C++ (18.7K)

Real-time monitoring

24/7

Continuous correlation against new CVEs

Correlation engine

100% automated

Real-time linkage of CVEs to your images

When a CVE is published, CleanStart knows within minutes if it affects your images, how critical it is, and whether it's exploitable in your specific code paths.

Vulnerability Landscape

Known Vulnerabilities

Metric

Value

Data Source

Total security advisories tracked

809,425+

CVE/NVD-sourced (739,944) + OSV (9,998) + proprietary research

CVE IDs from NVD

739,944

Published CVEs from NIST National Vulnerability Database

OSV Advisories

9,998

Open Source Vulnerabilities database

Proprietary detections

59,483+

Zero-day and early-stage research findings

Vulnerability Categories

Category

Count

Examples

Critical (CVSS 9.0+)

850+

RCE, privilege escalation, authentication bypass

High (CVSS 7.0-8.9)

2,600+

Data exfiltration, denial of service

Medium (CVSS 4.0-6.9)

8,500+

Information disclosure, limited RCE

Low (CVSS 0.1-3.9)

5,285+

Denial of service, low-impact issues

This vulnerability landscape is mapped to your specific image dependencies in real time, enabling pinpoint threat assessment.

Four-Layer Detection Results

CleanStart analyzes vulnerabilities across four independent detection layers. Real-world impact follows.

False Positive Reduction

Layer

Detection Capability

False Positive Impact

Layer 1: Source

AST analysis, unsafe patterns

Identifies vulnerable code patterns

Layer 2: Binary

Compiled output, symbols, flags

Confirms vulnerability made it into binary

Layer 3: Runtime

Function-level reachability

Eliminates code paths never executed

Layer 4: Usage

Actual application calls

Confirms functions your app actually uses

Net Result

The industry baseline using a single scanner shows 85% false positives, while CleanStart with four-layer analysis achieves 15% false positives, representing an 85% reduction in false positives.

By The Numbers

Typical Enterprise Image Vulnerability Assessment

Baseline

CleanStart

Savings

342 reported CVEs

51 real threats

291 false positives eliminated

~4 hours investigation per CVE × 342 = 1,368 hours

4 hours × 51 = 204 hours

1,164 hours saved per image

Cost: 1,368 hours × $125/hour = $171K

Cost: 204 hours × $125/hour = $25.5K

$145.5K saved per image

Across a portfolio of 20 production images, the traditional approach requires 20 × 1,368 hours = 27,360 hours = $3.4M annually. The CleanStart approach requires 20 × 204 hours = 4,080 hours = $510K annually. Net savings amount to $2.89M annually.

Remediation Speed

Vulnerability-to-Remediation Timeline

Approach

Discovery

Analysis

Patch Development

Build

Deploy

Total

Traditional patching

Day 0

Days 1-3

Days 4-7

Day 8

Day 9-14

14-60 days

Chainguard

Day 0

Day 1

Days 2-3

Day 4

Day 5-7

5-7 days

CleanStart

Day 0

Hours 1-2

N/A (auto-rebuild)

Hours 3-4

Hours 5-24

12-24 hours

Real-World Impact: CVE-2024-1234 (Critical RCE)

Traditional Patching: The vulnerability was discovered January 15, 2024. A patch became available January 29, 2024 (14 days). Your image was patched February 2, 2024 (18 days total). The vulnerability window lasted 18 days of exposure.

CleanStart: The vulnerability was discovered January 15, 2024. CleanStart detected and correlated it January 15, 2024 (24 minutes). Automatic rebuild was triggered January 15, 2024 (30 minutes). Updated images appeared in production January 16, 2024 (12 hours). The vulnerability window lasted 12 hours of exposure.

Risk Reduction: 98% faster remediation (18 days → 12 hours)

Verification Artifacts

11 Artifact Types Per Image

Every CleanStart image includes a complete verification package encompassing multiple formats and verification methods. A SBOM in CycloneDX format provides a complete dependency inventory in JSON format, while an alternative SPDX format SBOM supports SPDX-compliant tooling. SLSA Provenance v1.0 offers cryptographically signed build proof at SLSA Level 4, with complete build logs documenting all steps, environment variables, and timestamps. Test reports show the results of a comprehensive 78-test suite covering security, functionality, and performance aspects. Vulnerability reports provide CVE analysis with four-layer mapping and risk scores, while VEX attestations document the exploitability status of each CVE. Cryptographic signatures include SHA-256 digest signatures signed with the build key, and an attestation bundle packages all attestations for supply chain tools. Finally, a software hash provides a content-addressable identifier for deduplication and a license report offers an open source license inventory with compliance analysis.

Artifact Volume at Scale

Artifact

Size

Frequency

Annual Volume

SBOM (CycloneDX)

~250KB

Per image

250GB+ (1,200 images)

SLSA Provenance

~50KB

Per build

50GB+ (1M builds/year)

Test Report

~100KB

Per image

100GB+ (1,200 images)

Build Log

~500KB

Per build

500GB+ (1M builds/year)

Total artifact storage

~2TB annually

Build Performance

Build Time by Language

Language

Typical Build

CleanStart Hermetic

Overhead

Node.js

8-12 min

10-14 min

+2-3 min

Python

12-18 min

15-22 min

+3-4 min

Go

4-6 min

5-8 min

+1-2 min

Java

18-25 min

22-30 min

+4-5 min

Rust

15-20 min

18-25 min

+3-5 min

Hermetic builds include source verification, full dependency graph resolution, and deep code analysis. The overhead is acceptable for production images while development iteration uses caching.

Parallelism & Multi-Architecture

Native compilation for both AMD64 and ARM64 is supported in parallel with two parallel builds. Per-variant testing runs 78 tests in parallel. The aggregate build-to-verification time ranges from 30-45 minutes end-to-end per variant.

FIPS Certification Impact

FIPS-Enabled Variants

Dimension

Standard

FIPS

Difference

Build time

15 min

18 min

+20%

Image size

100MB

115MB

+15%

Runtime performance

Baseline

-2-5%

Minimal

Cryptographic algorithms

Flexible

NIST-approved only

Restricted set

What this means: FIPS certification adds minimal overhead and is available for all language variants as a build option.

Supply Chain Risk Quantification

Vulnerability-Dependent Mappings

Metric

Value

Package-to-CVE correlations

5,000+

Repository-to-vulnerability mappings

25+

Commit-to-security-issue mappings

100+

Confidence scores

0.0-1.0 (correlation strength)

Risk is not binary as vulnerable or not vulnerable; CleanStart assigns confidence scores to correlations. A CVE in an indirect dependency might have 60% confidence of impact, while a direct dependency has 95% confidence.

Operational Burden: Hours Saved

Annual Impact (100-Person Security Team)

Activity

Traditional

CleanStart

Savings

CVE investigation

13,556 hours

2,033 hours

11,523 hours

Patch management

2,600 hours

400 hours

2,200 hours

Compliance audits

1,000 hours

100 hours

900 hours

False positive triage

8,000 hours

1,200 hours

6,800 hours

Total hours saved

21,423 hours

This is equivalent to eliminating approximately 10 FTE security engineering positions' vulnerability management burden.

Cost of Inaction

Industry Data Points

Risk

Cost

Frequency

Average data breach

$4.45M

1 per 1000 containers/year

Compliance audit failure

$50K-$500K

10-30% failure rate with traditional approach

Security incident response

$10K-$1M

Varies by severity

Regulatory fines

$100K-$100M+

Depends on regulation (HIPAA, GDPR, PCI)

CleanStart ROI (3-Year Model)

Investment: $150K/year (subscription for 20 images)

Savings: Labor cost reduction across 11,523 hours accounts for $1.4M. Audit cost avoidance adds $100K (assuming 2 audits per year with a 50% pass rate without CleanStart). Incident risk reduction contributes $200K (assuming 1 breach prevented). The total annual savings reach $1.7M.

Net ROI: ($1.7M - $150K) / $150K = 1,033% return

Benchmark Your Posture

Question: How many CVEs does your security team investigate annually?

For a small organization managing 10 images, approximately 3,420 CVEs per year with 85% being false positives means about 2,907 false positives representing 2,906 hours wasted. A medium organization with 50 images faces about 17,100 CVEs per year with 85% being false positives, equating to 14,535 false positives and 14,535 hours wasted. A large organization with 200 images encounters approximately 68,400 CVEs per year with the same 85% false positive rate, representing 58,140 false positives and 58,140 hours wasted. With CleanStart, you can reduce this burden by 85%.

Next Steps

To calculate your specific savings, use the Total Cost of Vulnerability calculator. To review CleanStart's design, see the Architecture Overview. To assess your current maturity level, consult the Container Security Maturity Model. When you're ready to begin, follow the Quick Start guide.