SunRay Update 6

“You never get a second chance to make a first impression.”

The core of the new sunray release was to ensure the sunray server software were setup in an optimal manner. Moving onto new and powerful servers, with a fresh install configured for performance.
However the key to accessing this functionality is the menu screen. To focus to the latter stages of the project the menu has undergone several stages of development using the Waterfall model for software life cycle.

Obviously this element to the project is where the users will get their first impressions, deciding on their likes and dislikes of the system as a whole, so getting the front end as user friendly as possible was key to the design and planning. Research was conducted on several possibilities and in the end we narrowed it down to two potential languages; Tcl and Java. The menu which had previously been made and used was written in Tcl and with its two buttons, chiefly acted as a floodgate for the sunrays – if the menu were not there; the sunray devices simply cycle continuously waiting for a server or else connect to a terminal server utilising resources when there is no demand. But now we needed more functionality and implemented in such a way that actually reduces the hassle of logging in, rather than adding another stage to the process.

The prototype for the system seemed to perform effectively, although by no means complete it had the core functionality and no major issues arose. The most negative feedback was due to the slightly overwhelming blue and clumsy buttons.

Ideally we needed a simple login menu, without too much information, yet intuitive enough to understand immediately without guidance. Our original two button interface had roughly a dozen variations for connecting to different servers which made updates laborious. There must be facility to input a username and password, select screen resolution and be resistant to any tampering.

Given the adaptability of Java, it seemed like a logical move to create a visually superior model. Using java swing we created a slick and fully functional menu.

The JFrame itself is expandable, allowing for additional options to be displayed without requiring a new window.

However when analysing the memory overhead, the use of a Java menu utilised substantially more memory than the Tcl menu. This overhead unfortunately meant it would not be a feasible option when compared to the lower requirements of Tcl.
Thus we revert back the previous possibility of a Tcl menu, though there were several changes which needed to be made. The visual aspects of the menu were simplified and colours were swapped for a more sensible greyscale.
With the new buttons, a dropdown menu was created to reduce the requirement for extra shell scripts. Now the majority of the menu and its functions will run directly from the original Tcl script.

Finally, the only issue left with the menu was the slightly blocky fonts used. We had originally overcome this problem by using images rather than text labels, but this isn’t the ideal solution as changing text becomes an annoyance. Instead by downloading the most recent version of Tcl (and wish), smooth rounded fonts are available and the final result – I think – looks stylish and more importantly works efficiently and effectively.

This entry was posted in Cool Stuff, General, Green Computing, Sunray and tagged , , , , , , , . Bookmark the permalink.

5 Responses to SunRay Update 6

  1. Given the title of this blog post, any hints on what is coming in the future for SunRay users?

  2. Adam says:

    …the future is here and now!

  3. Andrew says:

    Isn’t the concept of decentralised (dumb terminal) computing an old technology?

  4. Adam says:

    Yes, the concept (and even use) has been around for years. Thin clients were originally called “graphical terminals” and an early example was the X terminals during the 90’s.

    However until more recently it has always been a struggle to ensure implementing such a system is not cost prohibitive.

    The cost of the servers themselves was always a major drawback in the past; to obtain a server powerful enough to host dozens of sessions (running seamlessly with the expected operating system), the thin clients and the appropriate licensing.

    Even today with all these factors, the immediate cost difference is fairly minimal, the big savings are in the long term – the reduced support, noise, maintenance, replacement time and costs, stability, administration and of course the power savings.

  5. Hi Andrew,

    It is an old concept but the SunRays use the latest technology. Dumb terminal were used widely by the first IBM mainframe computers and were text based terminals. PCs became more popular than dumb terminals when the prices of PCs dropped as they offered a number of advantages over dumb terminals including superior graphics and sound capabilities. The problem with PCs now is that they are unsustainable as energy costs rise and businesses are under pressure to reduce their carbon emissions. The SunRays use Windows 2003 (or Windows 2008 in the future) and have graphics and sound capabilities. There’s a really useful paper on this subject by Surachet Tanwongsval from the SANS Institute, it covers the history of dumb terminal / thin clients and the benefits of this type of technology. You may not have time to read the entire paper but chapters 3 and 4 deal specifically with this subject.

    Tanwongsval S, 2002
    Sun Ray™ Thin-Client and Smart Cards: An Old Concept With New Muscle

Leave a Reply

Your email address will not be published. Required fields are marked *