L3C Limited

L3C AI Deep Learning Cloud (IaaS)

Cloud infrastructure for AI, machine learning and deep learning. Begin with Proof of Concepts to prove your models then select the most appropriate deployment (on premise, cloud). Built on latest NVIDIA DGX Tesla GPU technology and/or IBM PowerAI. Optimised for high resolution images and large data models with 24x7 support.


  • Integrated deep learning frameworks and libraries
  • Optimised for processing images
  • Optimised for large data models
  • Scale without performance degradation
  • 24 hour SLA
  • Easy on boarding
  • At least 20% cheaper than AWS
  • Support and guidance from experienced technical team


  • Easy start for your AI projects
  • No need to purchase expensive infrastructure
  • Process larger matrices, higher definition images and larger batch sizes
  • Access latest technology
  • Fully supported 24x7
  • On boarding assistance and guidance
  • Highly cost effective solution compared to public clouds


£315 per unit per week

  • Education pricing available
  • Free trial available

Service documents


G-Cloud 11

Service ID

1 3 4 1 2 2 6 4 3 9 3 0 9 7 5


L3C Limited

Jerry Crossfield



Service scope

Service scope
Service constraints No
System requirements Linux - provided by L3C as part of the service

User support

User support
Email or online ticketing support Email or online ticketing
Support response times Priority 1: 15 minutes
Priority 2: 60 minutes
Priority 3: 8 hours

These can be adjusted on request
User can manage status and priority of support tickets Yes
Online ticketing support accessibility WCAG 2.1 A
Phone support Yes
Phone support availability 24 hours, 7 days a week
Web chat support No
Onsite support Yes, at extra cost
Support levels L3C provide a 24x7 SLA even for Proof of Concept environments.

Priority 1 response time is 15mins;
Priority 2 response time is 1 hour;
Priority 3 response time is 8 hours.

Telephone support is available 24x7

The Cloud service SLA is to exceed 99.5% availability
Support available to third parties Yes

Onboarding and offboarding

Onboarding and offboarding
Getting started Online web/conference call is used to enable users to access the service.
Service documentation Yes
Documentation formats PDF
End-of-contract data extraction User has full access so can download relevant data and then delete from the service and destroy the Linux container. L3C can also delete the data on customer request and we will provide a certificate to confirm deletion.
End-of-contract process At the end of the contract the user can copy their files and libraries and any data to physical media and request L3C to send to them (no charge).
They can also copy their files, libraries and data and download to their own systems and delete on the L3C Cloud environment.
They can also request that L3C store the data for a period of time for later retrieval or for use on say the next phase of the development project (Additional charge)

Using the service

Using the service
Web browser interface Yes
Using the web interface Using the OpenStack interface users can; launch an instance, boot from image or snapshot, connect to the instance using SSH, track usage of the instance, resize, pause or terminate the instance.
Web interface accessibility standard WCAG 2.1 A
Web interface accessibility testing We use OpenStack technology which has been extensively tested by the community
What users can and can't do using the API We use OpenStack APIs. Based on currently available OpenStack APIs, users can launch instances, create images and assign metadata.
API automation tools OpenStack
API documentation Yes
API documentation formats HTML
Command line interface Yes
Command line interface compatibility Linux or Unix
Using the command line interface We use the OpenStack command line. Installation and usage instructions are current provide at: https://docs.openstack.org/mitaka/user-guide/common/cli_install_openstack_command_line_clients.html


Scaling available Yes
Scaling type Manual
Independence of resources We use technologies such as Linux Containers to use a high degree of separation. Additionally physical resources such as GPU resources are assigned to the container, so that container can only use the physical GPU resource.
Usage notifications Yes
Usage reporting
  • Email
  • SMS


Infrastructure or application metrics Yes
Metrics types
  • CPU
  • Disk
  • HTTP request and response status
  • Memory
  • Network
  • Number of active instances
Reporting types Regular reports


Supplier type Not a reseller

Staff security

Staff security
Staff security clearance Conforms to BS7858:2012
Government security clearance Up to Baseline Personnel Security Standard (BPSS)

Asset protection

Asset protection
Knowledge of data storage and processing locations Yes
Data storage and processing locations United Kingdom
User control over data storage and processing locations No
Datacentre security standards Complies with a recognised standard (for example CSA CCM version 3.0)
Penetration testing frequency At least once a year
Penetration testing approach ‘IT Health Check’ performed by a Tigerscheme qualified provider or a CREST-approved service provider
Protecting data at rest
  • Physical access control, complying with another standard
  • Encryption of all physical media
  • Scale, obfuscating techniques, or data storage sharding
Data sanitisation process Yes
Data sanitisation type
  • Explicit overwriting of storage before reallocation
  • Deleted data can’t be directly accessed
  • Hardware containing data is completely destroyed
Equipment disposal approach A third-party destruction service

Backup and recovery

Backup and recovery
Backup and recovery Yes
What’s backed up
  • Files, databases, virtual machines
  • Configurations
  • Entire System
Backup controls L3C provide a standard backup schedule within the pricing model. A custom backup schedule can be agreed for an additional charge.
Datacentre setup Multiple datacentres with disaster recovery
Scheduling backups Supplier controls the whole backup schedule
Backup recovery Users contact the support team

Data-in-transit protection

Data-in-transit protection
Data protection between buyer and supplier networks
  • Private network or public sector network
  • IPsec or TLS VPN gateway
Data protection within supplier network IPsec or TLS VPN gateway

Availability and resilience

Availability and resilience
Guaranteed availability Minimum 99.5 % availability. If this is not met in a given month users are refunded 25% of the service charges for that month.
Approach to resilience This is available on request
Outage reporting Email alerts to users

Identity and authentication

Identity and authentication
User authentication
  • 2-factor authentication
  • Username or password
Access restrictions in management interfaces and support channels The user provides a list of authorised users to contact support / service desk and to receive service reports.
Access restriction testing frequency At least once a year
Management access authentication 2-factor authentication
Devices users manage the service through
  • Dedicated device on a segregated network (providers own provision)
  • Dedicated device over multiple services or networks
  • Directly from any device which may also be used for normal business (for example web browsing or viewing external email)

Audit information for users

Audit information for users
Access to user activity audit information Users contact the support team to get audit information
How long user audit data is stored for User-defined
Access to supplier activity audit information Users contact the support team to get audit information
How long supplier audit data is stored for User-defined
How long system logs are stored for User-defined

Standards and certifications

Standards and certifications
ISO/IEC 27001 certification Yes
Who accredited the ISO/IEC 27001 Transpacific Certifications Limited
ISO/IEC 27001 accreditation date 22/11/2018
What the ISO/IEC 27001 doesn’t cover N/a
ISO 28000:2007 certification No
CSA STAR certification No
PCI certification Yes
Who accredited the PCI DSS certification Our data centre provider Virtus is PCI compliant
PCI DSS accreditation date Can be provided on request
What the PCI DSS doesn’t cover We do not currently accept credit card payments for our cloud transactions. Our clients have preferred to control user access and therefore we invoice in arrears based on usage of the service (see pricing model)
Other security certifications Yes
Any other security certifications Cyber Essentials

Security governance

Security governance
Named board-level person responsible for service security Yes
Security governance certified Yes
Security governance standards ISO/IEC 27001
Information security policies and processes L3C are ISO 27001:2013 Information Security Management certified.

On a general basis the governance framework is based on:
• Understanding the criticality of information and information security to L3C and our clients
• Reviewing investment in information security in relation to our business strategy and risk profile
• At a Board level endorsing continued development and implementation of a comprehensive information security programme
• Regular reporting of the programme’s relevance and effectiveness:
Consequently, we:
• Define and review the policies and best practices
• Define a systematic approach to identify any risks
• Assign ownership and a Plan-Do-Check-Act model that analyses and assesses risk, identifies possible alternatives to mitigate and implement the solution

Information security is reviewed at Board level and considered a Board level responsibility.

L3C has a dedicated policy for risk assessment, which includes:
• Methodologies to calculate risk impact
• Plans for risk mitigation
• Risk treatment
L3C Information security team holds regular meetings to review security processes, risks, weaknesses and incidents. Items are raised at Board level as required.

Any actions are assigned owners and tracked to completion.

Operational security

Operational security
Configuration and change management standard Conforms to a recognised standard, for example CSA CCM v3.0 or SSAE-16 / ISAE 3402
Configuration and change management approach All the assets that make up the AI Cloud service are described and their configuration status held in our CMDB. Any configuration change is subjected to change management and the CMDB is audited every 6 months for discrepancies. We cover the lifecycle of all our assets from onboarding to offboarding and recycling.
A Change Advisory Board (CAB) supervises changes which are categorised as standard or emergency. The change process includes; steps to implement the change including order and dependencies; responsibilities; timelines and escalation procedures. Standard changes are proven and low risk, emergency changes are implemented asap.
Vulnerability management type Conforms to a recognised standard, for example CSA CCM v3.0 or SSAE-16 / ISAE 3402
Vulnerability management approach L3C utilises an annual penetration test performed by an external body CREST Approve body (BSI) and holds the Cyber Essentials certification. Additionally our Internal services are scanned and tested regularly for compliance with the security policies in the company to ensure we retain our ISO 27001:2013 certification.
Critical patches are deployed within 24 hours but our cloud infrastructure runs on IBM Power9 servers which are amongst the most secure on the commercial market place and NVIDIA DGX-1 servers. Therefore we receive our information about potential threats from vendors such as IBM, NVIDIA as well as Linux vendors such as RedHat.
Protective monitoring type Undisclosed
Protective monitoring approach We can provide our protective monitoring process on request to a potential user. If we find a compromise we implement our incident management procedure as described earlier. The response time will depend on the priority assigned to the incident and the associated change management procedure.
Incident management type Conforms to a recognised standard, for example, CSA CCM v3.0 or ISO/IEC 27035:2011 or SSAE-16 / ISAE 3402
Incident management approach We have pre-defined processes for security events, hardware failures and problems reported by our clients. Users can report incidents using a Priority Classification (1,2 and 3) to our service desk and each is responded to within a strict SLA. Incidents can be reported electronically via a service desk tool or by phone (24x7). Incident reports are provided to the user on an agreed basis (the standard is monthly but the frequency can be tailored on request). Our proactive monitoring may also identify incidents and similarly we will assign a priority to them and address accordingly.

Secure development

Secure development
Approach to secure software development best practice Independent review of processes (for example CESG CPA Build Standard, ISO/IEC 27034, ISO/IEC 27001 or CSA CCM v3.0)

Separation between users

Separation between users
Virtualisation technology used to keep applications and users sharing the same infrastructure apart Yes
Who implements virtualisation Supplier
Virtualisation technologies used Other
Other virtualisation technology used Our Machine Learning/AI Cloud infrastructure is built on underlying NVIDIA DGX and IBM Power9 and PowerAI technology. This uses Linux Containers as the virtualisation method.
How shared infrastructure is kept separate Production deployments will be on dedicated/private cloud environments. Proof of Concept or dev/test workloads are likely to be in a shared model. We use technologies such as Containers to ensure separation. In addition each container has a physical resources assigned to it such as GPU cards. Clients selecting the NVIDIA DGX AI cloud will have a dedicated environment.

Energy efficiency

Energy efficiency
Energy-efficient datacentres Yes
Description of energy efficient datacentres Our data centres follow the BREEAM standard and are BREEAM certified as VERY GOOD


Price £315 per unit per week
Discount for educational organisations Yes
Free trial available Yes
Description of free trial We provide the first week free of charge with support from our team to ensure the AI model is operational and the Proof of Concept or development or production deployment can begin.

Service documents

Return to top ↑