Monthly Archives: June 2012

I was sat on the plane coming back from Switzerland last night reading a variety of PDF's on my iPad I'd saved locally for research. The topic in my head was litigation in the Cloud and what the local and  international topics are around Cloud especially if you are contracting your Cloud presence to an outsourced third party.

In 2006 I launched litigation against one of the world's largest software companies and reached out of court settlement and payment of my costs and taxes, an exercise that I hope to never have to repeat. It started when one of their bloggers wrote an article on their load balanced geo-located mirrored public facing servers that then said something silly, inaccurate and litigious. It was a very large personal financial risk to launch legal action, a terrifying wake up call but successful in that they gave me a lot of money to go away and paid my attorney fees. I don't feel great that I won but it did teach me a valuable lesson in civility and in technology within the confines of the law. It also meant an exercise in cache scrubbing at the major search engines the costs of which must have dwarfed the damages I received - that I then donated in full to charity.

So having been through the pain personally of taking on an 800lb gorilla and having to personally fund legal action against their bottomless pockets it struck me that there is a new world order due right about now. When I launched my action, given I was taking on a company who have over $100bn of assets you had to have a good case and finding an attorney who understood the internet, storage, transmission as well as reputation and procedure.

I then sat on the plane with my wire-bound Red Hat notebook and looked at a typical service provider architecture taking into account key encryption and regular service level statements from providers I am aware of and started scrawling. One thing became very clear very quickly. If it had jumped off the page and illuminated the entire interior of the Airbus it wouldn't have been any brighter and more obvious.

Cloud is going to make wealthy legal practices even richer. Specialist paralegals and their research agents are going to enjoy a bumper pay packet in years to come and if I was a student studying for my law finals right now I'd be specialising in Cloud based eDiscovery and data collection methodology. A scant search of the internet results in not one academic institution globally who has picked up on this opportunity for specialisation. It will happen - you read it here first.

If you're serious about understanding service provision in the hosting arena and are thinking about drawing up service level agreements you first research exactly the liability of your provider cousins in the ecosystem. You then take independent advice from paid counsel and you print off a copy of the annually publised Fulbright & Jaworski 2011 Litigation Trends Survey, an independent survey of senior corporate counsel. You can read the 2011 edition here in PDF format - registration required. They break down litigation areas of growth and concern across business verticals such as Communications and Technology, HealthCare, Financial Services, Energy and Insurance etc etc. The one we're obviously interested in is Communications and Technology, there are both US and UK sections - I'm using the UK version below but the US version is just as well researched.

Once you've downloaded it - go read the section:

"Businesses may also face further issues in connection with new forms of data storage including 'cloud computing'. One in three (30%) UK respondents reported utilising cloud computing. One half (50%) of UK respondents using cloud computing have already reported having to preserve and/or collect data from the cloud in connection with actual or threatened litigation, disputes or investigations. More than one in ten (13%) UK respondents not yet using the cloud expressed an intention to move software or data to the cloud in the next 12 months.Worryingly, some 16% of all respondents did not know whether their companies stored data in the 'cloud'."

Then once you're done reading the Cloud / eDiscovery specific sections read Barry Murphy's excellent Forbes article on eDiscovery in Cloud. If you then do your homework  and look at how an Open Cloud compares architecturally to a more locked in layered Cloud architecture with abstraction layers and layers of complexity and retrofitted controls then you look at how CloudForms and using RHEV/KVM gets you to Cloud it becomes screamingly obvious.

A correctly built and architectured Open Cloud be it Private, Public or a blended Hybrid model does one thing properly - it is documented sufficiently to allow risk model mapping and data storage decisions easier on the part of the provider and the customer or the private customer using tools such as CloudForms to provision and deploy.It's certainly going to speed up eDiscovery and it's certainly not going to be as popular on an hourly fees basis for the attorneys charged with researching litigation claims in Cloud.With clear open shared API's such as Deltacloud (written by Red Hat and contributed to the Apache Incubator project), encryption, environments and architectures understood and documented it's going to be a reduction in one of the hidden costs of Cloud that nobody is daring talk about or even think about.

Being Open is good - it might just not be popular with the paralegals who are assuming an open checkbook when it comes to research and eDiscovery. By complete accident we gave birth to not just a better way of working in Cloud but we also reduced the future revenue growth of those who see Cloud as a cashcow waiting to happen.What can I say ?

We're just nice people making Cloud a secure capable place to put your work and doing it with transparency. We're getting those workloads to Cloud and we're doing our best to do it in a way that is industry friendly across international boundaries and makes sure people play nicely together.

CloudForms delivers a pretty big punch. The entire focus around delivering a properly defined open hybrid cloud across architectures specific to the needs of enterprise organisations is paramount. Let's just get that top dead centre. But to get to that stage in deployment safely you have to consider the application tier and all related operational capabilities and data. By thinking laterally, embracing CloudForms and realising that being rational that deciding a strategy around open hybrid design is the only way forward allowing you to reap the benefits of using existing applications without introducing costs, risks and issues.

It also helps you play nicer and be able to have that much more control than any other technology out there today.

Can you afford not to look at it ?

Addendum: If anyone can see pagination flow issues with this article I think WordPress have some serious issues in the last twenty four hours. Not a lot I can do as it's been tested in five browsers and the original article flows fine, it's just their posting API which is borked.

I had a great time in Switzerland meeting folk and giving a keynote at the Open.CH event in Bern. I've also noticed that Switzerland has jumped up to number 1 in page reads on this blog in the last three days outshadowing North American readers.

Seriously impressed with the smooth running but being Switzerland a country I have spent so much time in over my lifetime why am I surprised ? It's always nice to come back to the place I think of as my second home. A big player now in the Open world of technology advancement especially in the Swiss Government and home to some of the nicest people in the Linux world who I've worked with for what seems a lifetime but is still more than a decade.

I am probably going to be involved in a gig there in the Autumn and if you'll have me back next year then please promise me you'll all bring hunting rifles to hunt down the maniacs who drove round the square outside the railway station for four hours until 2am hooting their horns after the Spain - Croatia game. I got about two hours sleep and still managed to entertain and educate.

I saw a lot of unsure faces when I turned the microphone on but having watched the previous speakers delivering content I at least got a number of laughs and a lot of smiles out of a very educated and prodigious audience of my peers. I think it needed to be a little more light hearted and reach an audience sitting listening to slidedecks all day that are content and technology rich but don't often help you stay awake. My job is to keep you hanging on what I'm going to say next and have you feel we achieved something in the room. The fact that I have a load of new Swiss Twitter followers overnight shows you read the deck too which is always good news.

Thanks for having me and a special thanks to Matthias for his allowing me to keep you all amused and bemused in equal order. And thanks for the gift/award of the custom made Victorinox Swiss Army Cyber Knife. Since I was a small boy with my face pressed against the glass at the Victorinox shop in Bern I always wanted to take one home but was never allowed. Now I got to take one home aged almost forty years of age. It's much appreciated and not remotely expected. Hats also off to Philipp and the team at Red Hat Switzerland for sponsoring the event.

"Anybody can build a Linux server - even my grandmother could deploy a Kickstart image".

How many times have you heard that from a sysadmin / Linux admin ? Personally over the last decade I've heard it echo around. In 2012 we take it for granted that the work the community and the likes of Red Hat, SuSE and Canonical have bankrolled to ensure that the user experience for those taking their first steps away from other Unix platforms and Windows don't have the baptism of fire we did back in the day (I sometimes do hark back to building a Linux desktop and Xfree86 config burnt into my retinas and being able to get through those questions faster each time etc).

So whats the beef with the comment around Linux being easier to adopt and use and to consume as a free technology ? Well it's not really a beef but right now you find me in Switzerland sitting in another airport lounge typing this article having given a keynote on Open Cloud and engagement at a very well organised conference. Last night I met up with a few people locally in the Linux community at a higher level than most who have built offerings around Cloud.

All differed. All had their own pro's and con's, all of them had significant time and money invested in getting it right to build platforms for consumption by customers looking to move those all important paying workloads to Cloud.

All three had built out environments using Linux technologies, freely available Linux technologies. All three had another thing in common, they all started out in the 90s and early 2000's as ISP/Managed hosting providers. That realm of folk who were vital to the growth of hosting and email provisioning but a lot of whom hid behind proprietary licenced tools such as cPanel or BSD licenced technologies such as ISPConfig, Webmin or provided scripted provisioned access to tools such as phpMyAdmin for customers to be able to avoid getting under the hood so to speak. Thats where I'd say 70% of the managed hosting community in the EU and probably other territories got growth - by making things easy to consume.

It wasn't rocket science. It was basic web hosting, uptime and keeping disks spinning was your major issue - not the enterprise needs of customers. Customers with virtual webserver provisioned tiers typically didn't consume much support they generally only needed minor irregular assistance and you spent more time reading logfiles and racking hardware than you did thinking about how you needed to dovetail into your customers business needs and governance processes. In a lot of ways you could say pricing became a commodity product and many hosting providers added other services such as domain name registration, custom DNS and storage to their portfolios.

We haven't really moved as far as you'd think as an industry. So lets paint an analogy which might help accentuate the point I want to make here.

In the early / mid 1990s the major car manufacturers decided to add in bespoke proprietary methods of connecting specific car tuning and diagnostic gear to models at point of manufacture. So VW, BMW, Ford, Jaguar, Mercedes all had their own proprietary "plug" for a very short period of time to allow their service agents to be able to reset error codes and to diagnose issues with customer vehicles. Skyes Pickavant the motor aftermarket specialist started selling a box (that I helped develop the firmware for) to "read" the codes of all these disparate vehicles and a range of connections and leads to fit a variety of vehicles all with their "own plug". It was extending to the small dealer the ability to compete with the official service agent. Now almost every car manufactured since 1997/8 has an ODBII compliant interface (although some still need an adaptor). They all share a common serial command infrastructure for event escalation and identification but what it means is anyone can rack up to EBay and get a diagnostic reader that they can talk to their car with. I have one, it saves a fortune in fact I've lost count how many times other people have borrowed it.

Now what has this got to do with Cloud ?

The motor manufacturers were not open. They all had their own expensive R&D to do the same thing just using different methodologies. They were all trying to solve the same problem and wasting time and investment by not talking to each other and not sharing standards. In the end ODB II became a standard and everyone had to conform. In the same way that most phone manufacturers realise now using a micro USB connector is the norm as most devices use them to connect / charge.

Now these providers all made good money in the hosting community and branched out eventually to thinking about Cloud. However after listening to all of them, all having built what they thought was Cloud. I asked one question "now how are you going to attract and retain customers and by inference make revenue".

Silence. It's at this point I'm hoping I'm lost in translation and my question is being digested. It's clear nothing is coming in immediate response.

It's becoming clear. If you don't understand what your customer wants and needs and you don't make yourself a valued trusted partner with consumable elastic services and end point enterprise service level consumable architecture that ties into a contractually useful SLA then you should not be looking at Cloud. Stay in your comfort zone. It's like me deciding that tomorrow I'm qualified to do manicures and waxing because I can go to a cosmetic counter armed with a VISA card and a shopping list.

Thats not a rebuke. It's a statement that if you are going to stand up any commercial offering in any marketplace you must think about your business plan, your opportunities and risks and the tools you use to go to work.

If you buy cheap spanners and you have a DIY car manual don't expect a customer with a fleet of cars to want to sign up to a contract with you. However theres nothing to stop you working with an upstream provider or franchise (which is normally the route in fields as diverse as mobile catering to car dent removal) to make your business feel larger even if you have a major up front investment.

This doesn't really exist in IT, I was one of the first Microsoft Solution Provider partners (yes you read that entirely correctly) and it allowed us overnight to play in a bigger pool. In Cloud you need to do the same and you need to think about skilling up and getting a grasp of fast moving technologies in the IaaS and PaaS marketplaces as well as really taking on board my comments about cheap spanners.

There is nothing wrong with using a Free Linux OS such as CentOS. Just humour me and show me a religiously correct business plan for your CentOS built cloud which demonstrates actual retention of customers and growth. Theres a dynamic and significant reason why Red Hat long since started adding major synergies in the acquisitions of technology companies to help add to our stack. It's also the reason that we're seen as a progressive leader in Cloud.

Yes you can build yourself a basic Cloud using freely available community resources and yes you can support them yourselves. You could also build yourself a kit car and maintain it yourself but would you then enter a Formula 1 race with the kit car and no backup or pit crew or a factory and testing environment ?

Just because I can go to Home Depot and buy paint and brushes does not make me a trusted decorator. If you want to be serious about retention and attraction of customers to Cloud and you want to be skilled in every aspect of using Open technologies to get there safely, securely and to keep hold and grow customers then you need to be talking to me.

Red Hat are seasoned professionals not just at the operating tier of Cloud but also at helping you drive your business forward in order to attract customers and to retain and build revenue opportunities. It's a very open and intelligent ecosystem that is here to help you as a partner but please take it on the chin if in this article I've described your company. It's meant to make you think not point fingers at your architecture.

Engaging, Enthusing and Empowering Cloud - thats my job as an evangelist and I'm here as your resource so use me.

I arrived in Switzerland this evening to give a keynote on Open Cloud and Governance. After a day of driving, flying and trains I finally arrived at my tiny city centre hotel, a room on the top floor of a 1960s boutique hotel booked at a bargain rate (and now I know why). No aircon a window out onto the railway square plaza and main traffic area. It's all of 95 degrees and it's humid as hell.

Thankfully I am prepared and ready for the conference tomorrow. May record a podcast or two there and will definitely be taking pics to post here.

Will post updates and some new articles from there tomorrow if I get time.