sumo-tipping-point

Docker + Red Hat OpenShift = The Tipping Point for Open PaaS?

What if you could wrap up your appli­ca­tion in a light­weight con­tainer and then move it to any Linux server and have it per­form in a pre­dictable way with no changes? bare metal? vir­tu­al­ized? in a multi ten­ant envi­ron­ment? securely?

What if you could choose your  lan­guage and run­time envi­ron­ment (Ruby, Python, JEE, JavaScript, etc) and have it pro­vi­sioned with your appli­ca­tion code? with full life-cycle man­age­ment? with auto scal­ing? sup­ported by the ven­dor behind the tech­nol­ogy? on any PaaS?

TL;DR ver­sion — Docker is an awe­some way to con­tainer­ise any appli­ca­tion run­ning on Linux, but it has some short­falls. Red Hat’s col­lab­o­ra­tion with dot­Cloud the com­pany behind Docker will fix these and allow Docker run on almost any dis­tri­b­u­tion of Linux (includ­ing RHEL and it’s deriv­a­tives). They will also inte­grate the tech­nol­ogy that under­pins the Open­Shift PaaS (car­tridges) so that the life­cy­cle of an appli­ca­tion run­time can be man­aged. Mean­ing your appli­ca­tion will even­tu­ally be able to run any­where on your infra­struc­ture or on the cloud with no changes — cool eh?

Con­tinue read­ing

Checkmate, Nokia outwits Microsoft in the long game

When I saw the news about Microsoft Acquir­ing Nokia Devices and Ser­vices this morn­ing my ini­tial reac­tion was one of con­fu­sion, why would one loser acquire another loser? It seems that oth­ers had the same thought as well:
Twitter___ltm__The_Microsoft_Nokia_deal_...

 

How­ever as I digest this more I’m start­ing to think that this was part of a larger piece of chess where Nokia has realised for some time that the writ­ing was on the wall for their devices as Android and iOS were busy carv­ing up their for­mi­da­ble mar­ket share.

Nokia and RIM, who once dom­i­nated the smart­phone mar­ket have been in a steady decline for quite some time:

nokia_rim_market_shares_relative

It seems RIM’s attempt to get out of the funk was the Black­Berry 10 OS which hasn’t really worked out too well for them. Nokia on the other hand seemed to take a dif­fer­ent strat­egy, real­is­ing their ancient Sym­bian OS was a pile of dung they decided to do a an about-face and join with Microsoft and move all their phones onto Win­dows. Ear­lier this year this move was com­pleted and the last Sym­bian device shipped was in mid-2012.

So basi­cally Nokia killed Sym­bian, cut all the costs out of its busi­ness from rolling it’s own OS and focused on Win­dows devices and prof­itabil­ity of the devices divi­sion which it has achieved. They even sold off their head­quar­ters and leased it back! And the goal of all this — to inflate the price of Nokia Devices so that they could make a strate­gic exit by sell­ing to Microsoft.

This plan has clearly been in the mak­ing for a very long time and my hat goes off to what I think is a genius move by Nokia to offload a strug­gling divi­sion that once defined the com­pany so that they can once again rein­vent them­selves — which given they are Finnish and have a lot of Sisu I’m sure they will do.

Microsoft on the other hand have a lot of cash and I think this is a case where Balmer didn’t have much choice but to go along with this move, how­ever I don’t see a bright future for Win­dows phones — Microsoft needs to be in this mar­ket but they haven’t shown they under­stand it and they are too late in cre­at­ing an ecosys­tem. Maybe the next CEO of Microsoft will have a bit more insight and be able to take them some­where exciting?

Calculate the Business Benefits of PaaS

I’ve just noticed that over on openshift.com they have an intrigu­ing cal­cu­la­tor that claims to Cal­cu­late the Busi­ness Ben­e­fits of PaaS and to gen­er­ate a dol­lar value for mov­ing to PaaS, quot­ing directly from the page:

Plat­form as a Ser­vice (PaaS) is a pow­er­ful cloud solu­tion that can help solve real prob­lems today. With PaaS you can enable your devel­op­ers to work faster and allow your oper­a­tions team to pro­vide real busi­ness value by reduc­ing time-to-market for key IT services.

Use the Open­Shift Enter­prise Ben­e­fit Esti­ma­tor to help you cal­cu­late the poten­tial impact of:

  • Accel­er­ated appli­ca­tion development
  • Auto­mated appli­ca­tion pro­vi­sion­ing and config
  • Web-scale appli­ca­tion operations
  • Increased hard­ware uti­liza­tion efficiency

This tool can also help you esti­mate the ben­e­fits your orga­ni­za­tion could gain by imple­ment­ing Open­Shift Enter­prise in your environment.

Check it out here and give it a go.

I’d be inter­ested to hear how read­ers get on and it would be great if you could com­ment on this post with your results, anonymise your com­pany if you want, but it would be inter­est­ing to hear your thoughts on the results.

evolution_of_guybrush_threepwood

The Evolution of Paas and how this site was setup

As I’m a bit of a geek at heart and because I’m a major fan of open source I really wanted to host this web­site on an open source PaaS using as much open source soft­ware so I decided to use my favourite blog­ging soft­ware Word­Press and host it on Open­Shift Online. While I was putting this blog together I realised just how far things have come when it comes to set­ting up your own web­site, but before I dive into the evo­lu­tion of host­ing that has got­ten us to PaaS let me show you exactly why I think PaaS is so magic by tak­ing you through the three sim­ple steps it took to set up a Word­Press blog on the domain www.themiddlewareman.org.

Step 1. Reg­is­ter the domain (I use Gandi). This took just a cou­ple of min­utes includ­ing pay­ing by credit card.

Step 2. Setup an Open­Shift “app”. I used a word­press quick­start which only required the fol­low­ing steps and took a cou­ple of min­utes in total:

Cre­ate an account at http://openshift.redhat.com/ and install the client tools (run ‘rhc setup’ first)

Cre­ate a php-5.3 appli­ca­tion (you can call your appli­ca­tion what­ever you want)

rhc app create wordpress php-5 mysql-5 --from-code=https://github.com/openshift/wordpress-example

That’s it, you can now check­out your appli­ca­tion at:

http://wordpress-$yournamespace.rhcloud.com

You’ll be prompted to set an admin pass­word and name your Word­Press site the first time you visit this page.

Step 3. All that was left was to reg­is­ter an alias in Open­Shift for www.themiddlewareman.org and update my dns set­tings (I use DNS Made Easy) to point to my new appli­ca­tion url. Again this took only a few min­utes to do and as I’m using a really great DNS provider my changes prop­a­gated super fast!

Fun­nily enough tweak­ing the theme and con­fig­ur­ing plu­g­ins took WAY more time than instal­la­tion, but again the sim­ple steps above just illus­trate how easy it is to set up a web­site with a PaaS — and of course with Open­Shift I still have ssh access to the site and every­thing is in a git repos­i­tory that I con­trol and I can even do my own back­ups of my site if I don’t trust the Word­Press backup plu­gin I’m run­ning which backs up this site to my Drop­box account once a week.

So this prompted me to think back to just how far we’ve come in the last 20 or so years that the inter­net as we know it has been around. Here are the dif­fer­ent phases that we’ve been through so far.

  • Inter­net Con­nec­tion — yep, you got an inter­net con­nec­tion (rang­ing from a dial-up to a T3 line) and a sta­tic IP and you hosted your own site on your own hard­ware. This was pretty expen­sive and totally man­ual because you paid for Inter­net, Hard­ware and you looked after every­thing from hard­ware, soft­ware, net­work­ing and every­thing in between.
  • Servers at the ISP — either col­lo­cated or access to the ISP’s own servers. At this point you needed less net­work­ing smarts and your con­nec­tion got both faster and cheaper but you still needed expen­sive hard­ware and it was all still very expensive.
  • Geoc­i­ties arrived on the scene — It’s really worth men­tion­ing these guys because what they did was setup some of those big expen­sive servers and paid for the inter­net con­nec­tion but they carved it up into lit­tle chunks (1MB seemed huge at the time) and allowed you to host very sim­ple web­sites for free, they even gen­er­ated the HTML for you so you didn’t have to hand craft it your­self. This then spawned a whole host­ing indus­try where you could get a very small slice of a big­ger machine with vary­ing lev­els of priv­i­lege from what was usu­ally FTP and flat HTML files to SSH and the abil­ity to run CGI scripts.
  • Vir­tu­al­iza­tion and the arrival of the Vir­tual Pri­vate Server (VPS) — With the increas­ing usage of vir­tu­al­iza­tion by host­ing providers and as Linux started to infil­trate the host­ing mar­ket it became pos­si­ble to have your own vir­tual server that you had root access to and com­plete con­trol of. At this point peo­ple were really still build­ing tra­di­tional appli­ca­tions and man­ag­ing their servers by hand or with elab­o­rate man­age­ment tool­ing but the. Most of my own per­sonal web­site and projects are now hosted on a VPS that I’ve had for the last few years from Lin­ode (arguably they are an IaaS provider too) but with the evo­lu­tion of PaaS it’s start­ing to make more sense for me to move all these sites onto a PaaS.
  • Infra­struc­ture as a Ser­vice (IaaS) — The rev­o­lu­tion of IaaS was not that you could get a vm really quickly though that is part of it, for me the really sig­nif­i­cant change was that you started to archi­tect your appli­ca­tions dif­fer­ently, this is where appli­ca­tion nodes started to become state­less, we started to move away from clus­ter­ing to share appli­ca­tion state and move to using things like data­bases more for state (ini­tially, because that’s what we knew) and ulti­mately to using data grids for appli­ca­tion state. As appli­ca­tions get adapted or new ones writ­ten for this new approach they will become more resilient and scal­able with the main idea being that you can kill any server in your infra­struc­ture and ser­vice will con­tinue as nor­mal auto­mat­i­cally. With this approach you still do care about things like the oper­at­ing sys­tem and how to scale and all the magic that goes along with it. There are some steps being taken to ease this pain already in the IaaS world such as Amazon’s Cloud­For­ma­tion or the open source equiv­a­lent for Open­Stack called Heat which look promising.
  • Plat­form as a Ser­vice (PaaS) — this is really an evo­lu­tion of IaaS where instead of me (the appli­ca­tion devel­oper) main­tain­ing the under­ly­ing infra­struc­ture and tak­ing care of the oper­at­ing sys­tem, fail­ures and scal­ing etc. I now let the PaaS do that so that I can focus on my code and my appli­ca­tion alone. The steps to setup this blog are a pretty good exam­ple of how PaaS works.
  • Soft­ware as a Ser­vice (SaaS) — basi­cally hosted appli­ca­tions such as SalesForce.com or ser­vices pro­vided by Google such as Google Mail. I could have even gone with the SaaS approach by using a ser­vice like WPEngine to host this site and they would have taken care of all the has­sle or run­ning it and keep­ing it secure (at a cost). SaaS style appli­ca­tions are sort of what I expect most appli­ca­tion devel­op­ers to develop going for­ward and the smart ones will start build­ing their apps on a PaaS — like the Cloud9 IDE which is entirely run on the Open­Shift PaaS.

You may have more thoughts and ideas on these phases (com­ments wel­come) but for me as a lazy devel­oper who still likes to have con­trol I think it’s fan­tas­tic that I can now have an appli­ca­tion up and run­ning with a few key presses or clicks and what’s even more amaz­ing is that it’s com­pletely repeat­able and doesn’t depend on a huge IT depart­ment to run, any­one can host their own Open­Shift instance using either the Ori­gin com­mu­nity ver­sion or the fully sup­ported and tested Enter­prise ver­sion or just use the Online ver­sion if you are like me and don’t want to think about man­ag­ing servers anymore.

Is RSS dying?

So today I open up my RSS reader (I migrated to News­blur after Google killed Reader) and I’m brows­ing my usual feeds — includ­ing Dil­bert when I see this:

_2_129__NewsBlur

Grrr! Even after check­ing the web­site and going to the offi­cial feed it’s still bro­ken and the worst is that Scott Adams has decided to do this on pur­pose in some sort of per­verse and mis­guided attempt to drive traf­fic to their website.

You can’t beat the inter­net and geeks though and of course now there are alter­na­tive feeds like this one from Kimmo Suomi­nen which com­pletely fix things and made the Dil­bert move seem even like a move by the pointy head boss because it’s been worked around already.

Seri­ously though, why is such a great open stan­dard like RSS not being respected more, is it because every­one uses Face­book for their updates and it’s only us geeks still using RSS?

JBoss World 2012 Abstract Submitted

I just sub­mit­ted an abstract for a talk at JBoss World 2012, hope it gets accepted!

Real World Soft­ware Devel­op­ment with Openshift

Open­Shift is Red Hat’s free, auto-scaling platform-as-a-service for Java, Ruby, PHP, Perl and Python applications.

In this prac­ti­cal talk, Jeremy Brown will demon­strate how to use Red Hat’s Open­Shift PaaS within a team of devel­op­ers to rapidly develop software.

The talk will describe how a team of devel­op­ers can develop, test and release code in par­al­lel with­out los­ing devel­op­ment veloc­ity and will draw on real world exam­ples and insights.

Jeremy will demon­strate how to use and con­fig­ure a Jenk­ins Con­tin­u­ous Inte­gra­tion server with a Java project hosted on github.com, while mul­ti­ple devel­op­ers com­mit code and fea­tures in par­al­lel, with real time sele­nium test­ing as code is released. This talk will include code and con­fig­u­ra­tion sam­ples as well as a hands on demon­stra­tion of the end-to-end project workflow.

You will come away with a greater under­stand­ing of how you can use Open­Shift to cre­ate a highly effec­tive cloud devel­op­ment envi­ron­ment with zero server configuration.

Emergency Conference Speaker Required!

I had the slightly dubi­ous priv­i­lege this week of stand­ing in at the last-minute as a speaker at a Mil­i­tary con­fer­ence on secu­rity in Prague.

On Tues­day at 4:30pm I found out that we needed some­one to speak at a con­fer­ence in Prague the next day… slightly fool­ishly I vol­un­teered and 24 hours later I found myself presenting!

The con­fer­ence was the “Inter­na­tional Con­fer­ence ITTE 2011 — Cyber Secu­rity and Defense” - http://www.afcea.cz/ My topic was “Open Source and Secu­rity: Engi­neer­ing Secu­rity by Design’ and I had to brush up a bit on the con­tent as it’s not some­thing I had ever pre­sented on before.

Still it went OK I think and I enjoyed my brief 24 hours in Prague — though not to the 24 hours before sweat­ing and prep­ping for the talk and the panel dis­cus­sion that fol­lowed afterwards.

Here are my slides: Open Source and Secu­rity: Engi­neer­ing Secu­rity by Design