Concept

Rapid.Space provides public cloud and private cloud service based on Open Compute Project (OCP) [RD] open source hardware and SlapOS [RD] open source software for cloud automation. Every node of Rapid.Space is identical. A typical node consists at least of:

• dual Xeon (20 core, 40 vcore);
• 256 GB RAM;
• single 4 TB SSD;
• 10 Gbps LAN;
• 1 Gbps Internet transit shared by at most 36 nodes.

Rapid.Space service includes:

• hardware with maintenance;
• hosting, electricity and networking;
• cloud platform operation and support.

Rapid.Space cloud platform supports provisioning, monitoring, accounting and orchestration of:

• high performance virtual machines;
• high performance database clusters;
• global content delivery network (CDN);
• arbitrary custom services defined through devops scripts.

Rapid.Space is hosted in colocation datacenters:

• France: Data4;
• Sweden: Hydro66
• China: China Telecom (Xinzhou), China Unicome (Jinzhong) 

More nodes are planned in Russia, USA, Japan, Germany, etc. 

Rapid.Space nodes can also be hosted on premise to form a hybrid cloud solution.

Migration to Rapid.Space

Rapid.Space provides libguestfs [RD] open source migration tool from from physical machines (P2V) or VMWare/Xen virtual machines (V2V) to Rapid.Space high performance virtual machines. Migration can be done online if sufficient Internet bandwidth is provided. Migration can be also done on premise by bringing OCP servers to the existing data-center, doing migration on-premise and shipping servers to Rapid.Space data-center by courrier.

Rapid.Space offers optional migration services from existing infrastructure to its cloud infrastructure with the support of a network of partners. Current partners are located in: France, Germany, Japan, China, Bulgaria, Russia, Brazil, Italy. The partner network is growing.

The average effort of migrating and doing basic system level tests of a single machine is 2 hours or 1,600 RMB if this service is provided by Rapid.Space.

Recycling Hardware

Rapid.Space provides hardware recycling services whenever a company decides to migrate its legacy datacenter to Rapid.Space high performance virtual machines. Rapid.Space can purchase, recertify and reuse directly:

• Intel Xeon Scalable CPUs compatible with Tioga Pass platform [RD];
• Intel Xeon E5-2600 v3 (LGA2011-3) CPU compatible with Leopard platform [RD];
• Registered DDR4 RAM compatible with Tioga Pass platform [RD];
• Registered DDR4 RAM compatible with Leopard platform [RD];

Other parts will be sent to circular economy operators such as ITRenew [RD].

Thanks to recycling, part of the investment in legacy datacenter can be saved.

Cloud Services

Here are the different cloud services provided by Rapid.Space. Every node of Rapid.Space can be used in five different ways:

• Big VM with high performance disk I/O;
• Resilient VM with scalable disk, RAM and CPU;
• Resilient database cluster;
• Resilient PaaS;
• Content Delivery Network (CDN).

The table below summarizes the advantages and inconveniences of each type of service.

Big VM

Big VMs are ideal to deploy arbitrary software (Linux, Windows, BSD, etc.) which requires a lot of RAM, a lot of cores, fast LAN and maximum disk I/O performance. High availability can be implemented by the system administrator of the guest OS level with technologies such as drbd [RD] or by the application.  Disaster recovery must be implemented by the system administrator of the VM guest OS or by the application.

Typical applications include:

• Large application with database (ERP, CRM, etc.);
• Big data cluster (HADOOP, Jupyter Notebook);
• Business intelligence (Power BI, Business Objects, etc.); 
• Virtual desktop for CAD applications.

Read performance of SSD I/O of Big VM is 100% of bare metal. Write performance of SSD I/O of Big VM s at least 70% of bare metal. This is 2 to 8 times more than traditional virtualised storage [RD].

Resilient VM

Resilient VMs are small virtual machines that use virtualized storage and can scale dynamically:

• number of cores;
• size of RAM;
• size of disk.

Backup, archival and restoration of resilient VMs is automatic. A daily disaster recovery test is launched every day.

Resilient VMs are ideal for:

• small applications;
• development environments;
• compute intensive task with little I/O;
• automating disaster recovery of sensitive data or applications.

Resilient VMs are not suitable for high performance database.

Resilient Database Cluster

Resilient DB Cluster provides a swarm of databases deployed on a cluster of bare metal to achieve maximum performance and scalability. Supported database are:

• MariaDB;
• MySQL;
• NEO;
• PostgreSQL (option).

Resilient DB Cluster combines the advantages of Big VM with even more performance and the dynamic scalability of Resilient VM. 

Resilient PaaS

Resilient PaaS is a web based development environment with a strong focus on devops and automation of the entire lifecycle of software: build, provisioning, configuration, run, monitoring, accounting, orchestration, disaster recovery. It is based on nano-container technology which ensures portability across different operating systems [RD]. It embeds a framework to run continuous integration tests and automatically enforce that all operations can be reproduced over time and space.

Resilient PaaS is compatible with languages such as: C, C++, python, java, golang, ruby, Javascript, FORTRAN, rust, etc. 

Pre-Migration: Investigation

Rapid Space cloud provide corporate cloud solution to any size of business. In order to evaluate the need of a potential corporate user, the following table is provided for collecting information on user's current system configuration. 

Application Name: _________________________

Download this table to fill and send to Rapid.Space Team.