Quru Image Server - dynamic imaging for web and print
The Quru Image Server (or "QIS" for short) is a dynamic image server, offering a fully colour managed, dynamic resizing, templated solution that will fit the needs of e-commerce websites, image libraries, publishers and many other businesses.
The image server hosts an image file repository, along with its own image processing engine, management database, and dedicated web server. It can be used on its own, or alongside a separate web site, to store, manipulate and display images on the web in a variety of ways. A built-in web interface provides repository browsing, reports, and system administration facilities. Software developers can also use the built-in API to use the image server as a back-end for creating automated imaging workflows.
Quru provides QIS on an annual subscription basis, for a low cost compared with other dynamic image servers at scale. Commercial support, initial setup and systems integration services are also available. To request more information email [email protected] or use the request a call box.
QIS hosts and serves all the images for international auction house Bonhams, at www.bonhams.com.
Image processing features
QIS uses the popular ImageMagick package for its core image processing engine, along with enhancements for the handling of RAW image files and PDF documents. Imaging operations can be combined together, saved as templates, and include:
- Cropping, tiling
- Vertical and horizontal flip
- Overlays / watermarking
- Sharpen and blur
- Image format conversion (e.g. PNG to JPG)
- Image colorspace conversion (RGB, CMYK, and GRAY)
- ICC / ICM colour profile conversion and removal (for print publishing)
- Stripping of image metadata (e.g. EXIF profiles) to reduce file sizes
- PDF conversion to and from images
The supported file types vary depending on the installed version of ImageMagick, but the following file formats are enabled by default:
- General image formats: bmp, dcm, gif, jpg, png, ppm, psd, svg, tga, tif, xcf
- RAW image formats: arw, cr2, mrw, nef, nrw, orf, rw2, raw, raf, x3f
- PDF / Postscript formats: eps, epsi, epsf, pdf, ps
Embeddable image viewers
Processed (or unprocessed!) images can be displayed or made available for download
on your web site in a number of ways. The simplest and easiest is to use QIS with
See the imaging guide for examples.
QIS makes it easy to deliver responsive images to a wide range of different devices, whereby large devices download a large and detailed image, while small devices download only a small, and perhaps cropped version. This is a crucial technique for making fast loading and low bandwidth mobile web sites. Again you can see the imaging guide for an example.
- An inline image viewer with animated zoom
- When zooming in, only the visible portion of the enlarged image is downloaded
- When zoomed in, the image can be panned using a mouse or touchscreen
- Optional image information pop-up
- Optional full-screen mode
- A function to launch a full-screen zoomable viewer when an image or a page element is clicked
- An inline gallery viewer, showcasing multiple images or a folder of images
- A function to launch a full-screen gallery when an image or a page element is clicked (e.g. to launch a gallery of all the images of a product)
- An image carousel / slideshow, showcasing multiple images or a folder of images
- Selectable slide animation or cross-fade animation
- A lazy-loading image library
- Images in the web page are not downloaded until they become visible
For the adventurous, the image server's flexible APIs make it possible build your own image viewers!
Image repository management
As well as serving up images, the image server acts as a central place for storing, browsing, and managing your entire image library. Management features available in the web interface include:
- File system browsing
- Compact list view or a grid view of thumbnail images
- Add, move, rename, and delete image files and folders
- Image details viewing, including embedded EXIF data and file change history
- Image publishing tool with dynamic image preview
- Image uploads via drag & drop or local file selection
- Image usage charts
- Chart the number of views, number of downloads, bandwidth used, processing time, and response times for individual images
- Server performance charts
- See the top images by most viewed, most processed, most bandwidth, and slowest responses
- Chart the server's (or the cluster's) overall number of image views, downloads, caching success, performance, CPU and memory usage
- The creation of image collections - selected images that are viewable on a web page and downloadable as a single zip file (currently this is named portfolios and is only available using the developer API)
- System administration
- Define image processing templates
- User and group management
- User access control and image permissions
The image server holds a database of users and groups (which act like roles in other systems). Users are only required for logging into the web interface or accessing the API, so at its simplest you only need 1 user account for browsing the repository or performing administration. User accounts can be taken from an LDAP or Active Directory server in a corporate environment. Groups have a list of members (users) and are used for defining access permissions within the application.
Groups are also used to set the level of access that internal users and the public have to your images. You can keep all of your images private, have all of them public, or anything in between. This is achieved by choosing a folder from the repository, and then a group, and setting what level of access is allowed. For each folder and group, the available levels are:
- No access
- View only
- View and download the original image file
- Plus change image attributes
- Plus upload new images
- Plus delete images
- Plus create sub-folders
- Plus delete sub-folders (full access)
Everything that QIS does can be controlled programmatically through its Application Programming Interface (API). This enables the image server to be integrated with other systems, and for scripts to be developed to perform automated imaging workflows. The functionality provided includes:
- Authentication (login) and access control
- Image processing
- Image repository browsing and management
- Portfolio creation and management
- Image uploads
- Image template management
- System administration
Design goals and philosophy
The image server was developed with a few goals in mind. A discussion of these isn't entirely necessary, but it may be useful to understand why QIS does things in certain ways and how it differs from other dynamic image servers out there.
Single source, multiple uses
QIS operates on the basis that one image file is uploaded, and multiple uses of that file are possible from then on. If you are serving responsive images on a web site, one image can be served in several different sizes and with different crops for the best appearance. If you host a PDF document on the server, you can present users with a preview of the front cover (or indeed any other page), without them needing a PDF viewer and without requiring a separate image file to be stored.
There is no duplication of image files on disk, which keeps the file storage simple, with less to go wrong and less wasted space.
A simple image archive
The image repository is stored as normal files and folders in a normal file system. If you have an existing folder of images on a PC or a server, you can point QIS at that folder and the images will be immediately visible in QIS's web interface and for publishing on the web.
The image server does not mind if files appear and disappear in the repository, it does not mind if you use FTP or SCP or RSYNC or a network file system to bring in files from other places and other systems.
Because it operates from a standard file system, there is no lock-in with QIS, you own your data, and you can easily take your images elsewhere. You can also use standard file managers and standard backup tools to manage your files and keep them safe.
None of this can be said for image services that operate in The Cloud.
Performance and caching
QIS makes use of an in-memory cache of generated images for achieving its performance goals. This design relies on the assumption that once you have a requested a particular image (applying some specific processing parameters), that the image generated will be needed again soon. This is usually true of web pages, where the same images (or a fixed number of variants of them) are repeatedly delivered to everyone that visits the web page.
The caching engine currently used is Memcached. This has some drawbacks, but it is fast, simple, and reliable. Used alongside the image server, Memcached stores images and data in memory up until a memory limit is reached. This limit needs to be tuned to be as large as possible while still leaving some working memory free for image processing and anything else running on the server. Once the limit is reached, the next item to be stored causes the least recently used item to be evicted from the cache. In other words, Memcached prioritises keeping hold of the most frequently used images and data.
If you generate images that are always unique, or will not be re-used in the near future, then QIS will still work for you but it will not be optimal.
On the other hand, when you request an image that has already been generated and is still in the cache, it typically takes less than 2 milliseconds for the server to check your access permissions and fetch the image.
To request more information, or have a chat with us about the image server or any of our other services, email [email protected] or use the request a call box.
Quru is a market leader in the technical development, deployment and support of Linux and open source solutions that help organisations to reduce costs and increase operational agility and capability. We have also developed multiple award-winning software solutions ranging from mobile phone apps to global enterprise systems. Quru is based in Somerset House on the banks of the Thames, right in the centre of London. More...