QuickCode for Corporates
QuickCode helps people manage datasets, and the tools needed to get, clean and analyse data.
This page describes the corporate version of QuickCode, covering matters relevant to legal, IT and financial staff. Levels of service for non-corporate level customers are lower, and are described in separate terms and conditions. End user documentation is also separate.
We undertake to describe all aspects of our service clearly and transparently. If you have any questions, please get in touch and help us improve this living document.
Type of cloud service
QuickCode is a hybrid of SaaS and PaaS. It provides end user tools like a SaaS, yet allows visibility and customisation of all source code like a PaaS. This is because data tools need to be easy for anyone to use in most cases, yet powerful enough for those who can program to handle edge cases when needed.
QuickCode is deployed in a Public Cloud model. It is hosted on IaaS servers at Amazon Web Services in Ireland. We meet Amazon's tiering standards.
We currently scale by provisioning as required, as we are a platform for collaboration on (relatively) small amounts of data. We outsource any very large data volumes to other providers. By default we offer non-guaranteed performance, we can offer guaranteed for a larger cost. Utilisation reporting, where there are restrictions, is built into the platform and its API.
Our platform can be freely used by our customers in any way they like. In particular, customers can contract third parties to adapt, extend or write new data tools and applications which run on the platform, as they want. Even the code of the platform itself is open source.
Languages / protocols
QuickCode is a web platform for data science. Subscribers to QuickCode can run scripts which get, clean and analyse data – such as scrapers and visualisations. It can be used to gather information from disparate sources, including websites, FTP servers, legacy database systems, PDFs, and then export that in a common format (Excel, JSON, CSV), or as interactive web visualisations.
- Any open source programming language – Python, R, Perl, Java and so on.
- Schema-light SQL datasets (based on SQLite or PostgreSQL).
- Integration between datasets and views using web-standards based SQL API, which returns JSON over a REST interface.
- Full POSIX filesystem.
- External integration via SSH, SFTP, SCP, Git, Subversion and other standard Unix tools.
Good to know! Each dataset is a full virtual Linux shell account, meaning you can run any other Unix software you like, including binary-only software (e.g. Abbyy's PDF reader).
Backups / disaster recovery
All dataset files and data are stored on Amazon Elastic Block Store (EBS) volumes. Metadata is stored in a MongoDB server on MongoHQ’s cloud service.
These are five sequentially worse disaster situations, with recovery times indicated.
|Storage hardware failure||We store all important data on EBS, which is highly reliable. Volumes are automatically replicated so they are not vulnerable to failure of any single hardware component.||No time|
|Single machine failure||We can automatically rebuild individual machines in about 20 minutes, and the whole cluster in under 2 hours. We mount the EBS image from the failed machine on the new machine, so no data is lost. MongoDB is backed up to Amazon S3, where we can restore in under 1 hour.||~1 hour|
|Filesystem or EBS failure||Should an EBS volume fail or if the files on it accidentally damaged, we will recover it from daily, rotating snapshots. These are kept in two separate Amazon regions. We can restore the whole filesystem in under 1 day, and corporate plan customers in under 2 hours.||<2 hours|
|Datacentre failure||Should the Amazon Ireland data centre we use fail, they have additional data centres which we can restore to in under 1 day for all users, under 2 hours for corporate customers.||<1 day|
Should you wish to leave
We try to keep your custom by offering excellent service, not by locking you in.
Your data and files are available programmatically at all times, via our APIs and using the SFTP file transfer protocol. You can use this to make your own backups, or to migrate data off QuickCode. There is no extra charge for this. Datasets and views are written using standard open source tools and protocols, making it relatively easy to migrate individual applications to your own Linux servers.
In addition, the QuickCode platform itself and the core tools are all open source (see our GitHub account). This gives you the ultimate protection of being able to host it yourself, or pay another organisation to host it for you.
Purge data! Should it be required for compliance or other reasons, data can be destroyed. This can be done at any time by administrators of the account, or by asking us to. It will remain in backups until they are rotated out.