 Thank you. Hi everybody! My name is James Peale, I am product manager for Opsvue, which really translates into me being a project lead for the Opsvue Software project. I thought I'd start at least throwing one out of time at the end. What I'm going to do after this presentation is so I wonder off to the bar behind the main Johnson train theatre. Yn amser ond, nid yw'n gallu cwestiynau, ddim i'n gwaith, yw'n gweithio ar y bar i ddod 5-5.30, ddim yn ddim yn ddefnyddio. Yn amser'n gweithio, rydyn ni'n gweithio ar y cwestiynau. Rwy'n cael ei gydag i'r ffild y gallwn arlasau, ar y cyllid hynny, rwy'n gweithio ar y cwestiynau, wedi tyfn i'n gwylliant. Sa'n ddweud, mae'n gweithio. Yn amser y gallwn yn gweithio ar y cwestiynau. Rwy'n gweithio ar y cyllid hynny. dyna'r gwahanol i'w ddigonol heddiw i'n meddwl. Mae'n meddwl i'r gwaith yn meddwl, ond dyna'r idea wedi bod y brosiect yn gyflwyno'n meddwl a'r gwahanol i'r meddwl. I wna'r gwahanol i'r meddwl, byddwch chi'n gweithio bod yn cyfnod o'ch gweithio i'r ffordd o'r hyn o'r gweithio a'r mewn gwahanol i'r meddwl, i'w ddweud i'r gwahanol i'r gwahanol i'r ffordd o'r dyfodd o'r meddwl. Opsu sy'n meddwl o'r cynhyrchu yma, dwi'n meddwl y bydd yn y cyfrif, i'r cyfrifau a'i cyfrifau a'r systef yn y cyfrifau ar y sylwg, ac mae'n meddwl o'r cyfrifau a'r cyfrifau a'r cyfrifau, mae'n meddwl o'r cyfrifau a'r cyfrifau a'r cyfrifau a'r cyfrifau. Fel cwmhaenuits oherwydd, y bod yn iawn y gwahod rhagorau adegai gyda erioeddol y rhagorau adegai, yn myfyrwyr ef amddangas yn unig iawn ychydig ar gyfer hynny maelod rhagorau adegai a'r cyfrannu. Fy llyn nhw'n myfyrwyr ein bobl yw'r newydd yn ychydig ar gyfer rhagorau adegai, a'r gwahod rhagorau ar gyfer hynny yn dwi wneud. Felly, yr rhagorau adegai yn ychydig ar gyfer rhagorau ac mae'n ddweud eich gwahod rhagorau adegai. We've integrated many open source components, and as you see at the bottom, Opsu is completely open source, so partly what Opsu does is brings together a number of other open source components all under one umbrella. Opsu is designed for simplifying the task of managing your monitoring system. The software is commercially backed by Opsura, so Opsura basically pays for the developer's time and effectively the main sponsor of the project, and as I said, completely open source. So what features does Opsu have? Firstly, as I said, it has a web-based tool for configuring and managing Opsu, so obviously I accept that maybe one in this room I'm sure is quite capable of manipulating the configuration files that Nagios uses, and Opsu isn't designed to sort of dumb down that process, really the reason for having this web-based framework is to make it more efficient and allow you to manage very large systems so it scales up. Just grab a sip of water, and it just really is there to make your job more efficient, rather than sort of saying, well, you're not capable of handling these config files, we're going to give you a web UI, so it's not about dumbing down at all. Another key feature is Opsu has distributed monitoring, so basically that allows you to monitor very large networks, maybe across multiple locations, so you can have your master server in one location, and then you can have slave servers in every single data centre, you can have clusters of slave servers, so those slaves can obviously monitor large numbers of devices in each location, and we have some very large systems being monitored by Opsu, and that's all managed from the central monitoring server, so you have a central view on what's going on, and you have a central place to manage everything. We have SNMP discovery and polling, so SNMP being the most widely used protocol for monitoring and managing network devices, so in Opsu we will do a certain amount of discovery, and we'll poll network devices, pull information out of them. We also have SNMP trap processing, so this is when the network devices are sending out information, pushing the information to the monitoring system, and we'll take those traps, aggregate them, and based on a set of rules decide what to do with them, so we'll look at what the trap says and then go right, well, actually for this we want to raise a warning alert and send that out to the network admins. We also have graffing of performance and trend data in Opsu, so you can kind of see over time what the utilisation is of your storage network, you can see what CPU load is over time, and we have a long term storage database with the data warehouse, which you can use to do some analysis and generate reports from. And as well as the web UI, we also have an API, or several APIs, one for configuration, we have an API for notification, we have Nagios's API for generating alerts, so the point about that is that you can actually automate the task of managing your monitoring system, it isn't just about using the web user interface, you can use the API to, you can script against the API to actually generate all the objects and fire that, which is very good when you've got sort of built environments or using virtualisation, you can hook it all into the same set of scripts. So why Opsu, or perhaps more accurately, why should you be interested in Opsu? As I said, it's scalable, it's very flexible, partly because it's open source and partly to do the way we've designed it. We like to hope it's easy to maintain, that's one of the main design considerations. It's mature, and I'll come on to the history in a minute, and it's Nagios compatible, so in one sense we're sort of benefiting from the fact Nagios is widely used and widely understood, but also we try and contribute back where we can, so we contribute a lot of patches upstream and a lot of modifications, many of which get incorporated into Nagios 3, for example. So in terms of what's in Opsu, Opsu's mainly developed in Perl, so we have a lot of Perl and CPAN modules. We use catalyst, which is a web framework, so the application is written in catalyst, and MySQL is a database setting underneath it. In fact we have four separate databases that comprise the Opsu system, one containing the configuration, one containing the current status information, have a data warehouse and a reporting database. We bring together Nagios and a bunch of associated Nagios tools or projects, so Nagios has its own ecosystem of projects, so we integrate a bunch of those together under the Opsu umbrella, and also bringing NetSNMP, MRTG and NMIS, and that's an order complete list, that's just some of the more significant components. So a bit of history and another sip of water. The project was started in 2003 and really it came out of working with Nagios and doing some projects around it and thinking well, or in fact the people we were doing the project for were saying Nagios is great but it'd be excellent if it could do these extra things and maybe it'd be nice to have a web URI to actually configure and manage the system. So that's what really started the Opsu project and it really was initially just a bunch of bolt-ons to Nagios, but then quite soon it became more than just a bunch of bolt-ons and a more integrated framework for allowing us to manage the configuration and pull together all these underlying components in a coherent system. So I mean it isn't just add-ons to Nagios, it actually is quite a large amount of code sitting around that now. So version one was released in 2004 and I'm still heavily involved in the development then. And then Tom Voon joined in 2005 and his first release was 1.5 and Tom is also the development lead for the plugins project, Nagios plugins project, so he's heavily involved in Nagios through that as well as Opsu, he has several roles. And then Tom sort of took on majority of development work into Opsu version two which was released in 2006. And since 2006 until the end of last year really we were continuing to develop Opsu version two, so 214.3 is the latest release of the two zero stream or two stream. And then towards the end of last year, actually in August in fact, we were out in San Francisco at Linux World and we won the award for best systems management tool out there which was great. And we're up against some of the other open source system management tools, so that kind of was pleasing for us I think and we weren't expecting it, so that was a good experience. And then from the recently about three days ago in fact we released Opsu version three, so that is just released. And that's so exciting for us and a good step forward and provides a nice sort of set of foundations for developing more functionality within Opsu three as we move forward. And I'll come on to that in a bit later. So this sort of touching on the commercial aspect of what we do and this isn't meant to the sales pitch, it's just sort of just to give you a bit of insight into how we work really. So Opsu sits within a commercial framework, it's part of a, there's a commercial company sitting on top of it, like there is in many open source projects now. And the company makes this money out of services, again, not an unusual model. And the core development team work at Opsuera. So basically Opsuera is sponsoring the development of Opsu and making money out of sort of providing services around Opsu, support being the most obvious one and training and implementation and so on. And we also get a lot of functionality within Opsu sponsored by users and customers. So probably 70% or so of the functionality within Opsu has been sponsored by customers, which is kind of quite a nice way of funding the project. And then the rest of the infrastructure and time is sponsored by Opsuera. We have an active community of users and that sort of seems to be growing and there's definitely some very active users. We've got several hundred people on the mailing list and I know we've got sort of hundreds of systems out there. I can't say we've got thousands but it's certainly in the hundreds the number that we do have and we have a number of customers. So I suppose the significance of that is that we're actively getting involved in these systems and we learn from working with customers and that goes back into the development. So what's in Opsu version three? The main chain is based on Magios three. We've got a new performance graffing framework within Opsu three so that makes it easier to get the performance graffing information out using RRDs and also stored in a better format. We've improved the data warehouse. We've made improvements to performance, including faster reloads, especially on very large systems. So doing the reload cycle is 10% quicker and we expect further improvements there. We've integrated Magvis which is another popular part of the Magios ecosystem and that's a nice tool for visualising sort of status information. We've included better out-of-the-box configuration so when you set the system up there's a better configuration set off with and made lots of minor improvements. In terms of the future, for Opsu version three we've developed a much more formal process for putting a roadmap in place. So we've got a roadmap covering the next 12 months and we know pretty much what we're going to develop within Opsu and I'll come on to that if I've got some time afterwards at the end of this. We have a larger core development team so there's more people working on the software and I expect that to get bigger as well. We've been improving our test infrastructure which is always kind of a challenge so you develop the code and then need to make sure you're testing it properly before you do releases. So we've improved it already for version three and we expect it to improve further. There'll be more add-ons as well so we already have add-ons like SMS gateways and helpless integration tools and some reporting tools and we expect those to get more developed over the next year as well. To find out more, so we have the Opsu.org website, we have a download site with all the packages and there's a VMware demo VM for you to download which basically you can boot up and has a working Opsu system in there. And we have our IRC channel and if you go to Opsu.org there's plenty of other places you can find information such as a documentation site and mailing this and so on. In terms of the roadmap which I haven't put on the slide because it's not quite sort of pinned down yet but just to give you a bit of insight of what we have on there at the moment. There are plans to redesign the user interface really in terms of sort of look and feel. I think it's functional at the moment but not terribly pretty so I think the plan is to make it look more appealing but obviously not take away from the usability and hopefully improve that in fact. We have plans to include multilingual support within Opsu at the moment. There isn't a framework for doing that but the plan is to put the framework in place and then we can start doing translations into other languages other than English. We want to have developed some more event-based views. At the moment, monitoring views are very much oriented around status what's happening now. Plan is to have event views, maybe what happened in the last six hours, what is the sort of correlation between all these events. We plan to have the ability to also start creating your own custom views built on the new graffing framework and the events views and the status views so you can actually create your own views. Better integration with an open source reporting framework. I think that's an area that we need to do some work on so we won't develop our own framework. We'll use one of the open source ones and provide better integration, probably some templates. We want to improve out-of-the-box support for cluster monitoring. Again, it takes a little bit of work to monitor clusters so HA, sort of Linux clusters or network devices with this sort of active passive setup so better out-of-the-box support for that. Also, we want to extend and improve our contact profiles so you have a bit more flexibility with that. That's some of the things we're looking to do on the roadmap. There's some other functionality that I can't sort of announce at this stage but we will be talking about it before too long. I think I've almost run out of time. There's any quick questions. I'll take them otherwise. I'll see you in the bar behind Janssen. Hi at the back. We're using NSCA with a whole load of our own code. It's a simple answer to that. There's two-way communication to the slave and we push the configuration out to the slave. The slave doesn't monitoring. I've run out of time but I'll have a chat later. Thank you everybody.