Prof C. Paul Bonnington

This afternoon I will be speaking at a workshop on advanced videoconferencing and collaboration being hosted by the Human Interface Technology Lab at the University of Canterbury, and sponsored by KAREN.

I’ll be providing an update on the tool called Enabling Virtual Organisations (EVO). And, I’ll be doing this ‘remotely’ from my office at the University of Auckland to an auditorium in Canterbury using EVO itself.

EVO is a desktop computer-based collaboration tool which features advanced videoconferencing, as well as a host of integrated tools to facilitate collaboration between researchers and educators. I often describe it to people as ‘Skype Video on Steroids’. Reading this, you might feel I am an ‘EVO’ salesman, but I have nothing to sell; I write as an enthusiastic user of a tool which has fundamentally impacted on the way I do collaborative research.

EVO is designed to cater for the researcher or educator who wishes to collaborate with groups of colleagues globally without leaving their office. The design of the tool is very clearly centered on end-user experience. It is aiming to give the control back to the users so they are not reliant on other facilities and support.

Few (if any) barriers

What I regard as the most important aspect of EVO is that there are few (if any) barriers to getting started and using the tool to its full potential. From an end-user point of view it is very intuitive and easy to use. Most crucially it is free (as in beer) and educators and researchers don’t need to belong to any special “club” or request permission to use it. Other than a regular computer, users require just a regular (Logitech) webcam and mic-headset. (And did I mention it is Windows, Linux and MacOSX friendly?) Users self-register from anywhere, and they feel in control of the whole process. Quite simply, for the average researcher or educator there is virtually nothing (outside possible institutional network configuration issues) stopping them from utilizing the tool from their office and being able to lead/organize their own collaborative activities with others.

BeSTGRID’s role

EVO was introduced into New Zealand by the project I direct called BeSTGRID, which establishing an eResearch ecosystem for New Zealand researchers. We haven’t widely publicised EVO, (preferring to work with a smaller number of exemplar projects) but there are already around 300 self-registered EVO users in New Zealand. This user-base has grown somewhat organically by word-of-mouth, which reinforces the notion that the system is easy-to-use with no barriers and welcomes any user. There is even a growing community of users from the New Zealand high schools sector that are part of the KAREN “offramps”.

BeSTGRID provides a comprehensive information site for getting started with EVO in New Zealand.

EVO scales

EVO meetings don’t have to be small events; many people can participate, whether it is from their own desktop running EVO, or a standard ( Polycom, Sony, Tanberg etc) video conferencing unit, or by regular telephone, (even 3G video phone using SIP). For example, we regularly use it for BeSTGRID steering committee meetings involving 10 or more simultaneously participants from their offices scattered across New Zealand.

Unlike other systems (such as standard video conferencing, desktop-based Polycom PVX etc) the video quality of the individuals doesn’t degrade as more and more participants are included. And there is a wealth of other tools to support collaboration alongside the videoconferencing component.

• the ability to share with the other participants any document, presentation, or whatever you see on your computer screen (i.e a web site). Conveniently, your mouse pointer is “exaggerated” to the other users so they can clearly see what you are pointing/referring to
• a shared whiteboard for brainstorming (which is particularly effective for users with tablet laptops, as I have)
• text chat (private, room-based, or public)
• the ability to have private one-to-one discussions with another participant without the two of you having to leave the meeting room
• integrated meeting video recorder

EVO Compared to Access GRID

I am a great supporter of New Zealand AccessGRID infrastructure, and believe that AccessGRID is the right tool for many situations. However, there are some areas where EVO can be more effective:

• EVO is not fundamentally reliant on multicast networking (which can be extremely problematic between offices on a global scale) for large meetings
• Virtual meeting rooms can be created on-the-fly by any participant - no need to book ahead
• The meeting rooms can be secured by a key, which is simply a password which you exchange with the other participants before the meeting - that is ,no more ‘gate-crashers’
• Users are typically engaging from their desktop, eliminating the need to have to book a physical facility
• Simple interface with all tools integrated from the start; simple equipment, removing the need for an operator and reducing the need for installation support

I believe many situations/meetings that are currently taking place on the New Zealand Access Grid, could more easily be conducted on EVO. Particularly when there are small number of people at each location, or people are having to travel distances to get to an Access GRID physical facility. Furthermore, EVO has been demonstrated to function perfectly adequately in some situations where the tendency may be to try to use Access GRID: for example in remote teaching and seminar presentations.

Unique globally distributed server architecture

EVO is supported by unique globally distributed server architecture consisting of a large number of interconnected-Panda servers spanning the globe. Information (user information, meetings rooms, bookings etc) are constantly being replicated across all of the Panda servers, and the video data is streamed between them and onto the end users.

The main advantage of this approach is that the underlying infrastructure is extremely resilient to faults, and it is totally transparent to the users. When a user starts the EVO client, they are automatically connected to the nearest Panda server (using unicast). All communication is routed through that Panda server and tunnelled on to the other Panda servers (if there are other users in the same meeting connected to different Panda servers.) If a particular Panda server goes down for any reason, affected users are transparently connected to the next nearest Panda server: all they notice is the slight pause in the meeting of a few seconds. They are also transparently connected to other Panda servers if there are network congestion issues or server overloading.

BeSTGRID deployed the two Panda servers for all of New Zealand (there are two so that there is adequate fail over). Panda servers exist in Australia, Asia, North and South America, and Europe. Even communities outside those regions can use EVO provided that they connect to the Pandas with sufficient bandwidth.

How is Evo developed and supported?

EVO was born and raised in the academic/research community. EVO is a new tool (officially launched June this year, but based on VRVS which goes back to the late nineties). The development of EVO has been funded by the US Department of Energy to create the virtual collaboration platform for all of the researchers throughout the world involved in the Large Hadron Collider project at CERN.

A number of groups are involved in the development, but primarily this is led by Caltech. (For example, the developer community even includes a number of people working for the Slovak Ministry of Education.) Other communities and domains outside of European-US high energy physics community are very welcome to use EVO - and it will always be provided freely to researchers and educators throughout the world. However to ensure that the global distributed server infrastructure is coordinated and sustained, there is a subscription-based support model being used for the infrastructure across all domains and communities.

New Zealand, through the BeSTGRID project, and Australia, through AARNET are currently receiving this support freely on a trial basis from the EVO project team - but in future we will be looking to pay a subscription for this crucial service. Caltech, and in particular the project lead Phillipe Galvez, have been extremely generous with their time and support and we have felt very welcomed into the EVO community. At the end of the day, everybody involved wants a sustainable and coordinated environment to ensure that EVO can continue to be developed and improved.

How has EVO advanced over the past 6 months?

This business model ensured that there has been some major updates since April this year, when I first spoke about EVO at a similar workshop:

• A meeting recorder (which can be used by any of the participants) which is able to record and playback all of the video streams (including any broadcasted documents and desktops) and audio and real time. Furthermore, it is able to capture dynamically any whiteboard activity, or text chat.
• Multicast will be utilised where it is reliable (EVO itself transparently determines this), thus ensuring that the size of the video conferencing sessions can scale as-well-as AccessGRID.

• And finally, what I believe is the most interesting development, is the incorporation of Shibboleth authentication to enable Federated Identity Management.
  • Long-term this is going to have a significant impact, particularly in the higher education sector across the world. Simply speaking, the enabling of Shibboleth authentication in EVO will mean that in the future, all staff and students in New Zealand universities and research insitutes (and of course across the world) will automatically have a login to EVO. They can simply use their home institution’s credentials (that is, login/password) to login to EVO. This will improve the trust fabric in EVO, since the participants you invite to your meeting have had their credentials verified by a’reputable institution’ confirming they are ‘who they say they are’, (every time they log on.)
• Plus there will soon be an Open API for third-party development of tools, and advanced community management tools for communities to manage their own Panda servers