Podcast: Play in new window | Download
Subscribe: iTunes | Android | RSS
Justin and Jason interview guest Patrick Foley, host of the Startup Success Podcast, about his background as a software developer and his love of startups, why and how he got started as a podcaster, his role as a Microsoft ISV Architect Evangelist, Jason Cohen’s new startup WP Engine, the benefits of Azure as a scalable web platform and the power of the story in conveying ideas and understanding.
Reminds me when I first listening to TechZing, I had the impression that draws similarities of dynamics between the hosts with the Startup Success Podcast: which I think of Patrick being similar to Jason and Bob Walsh to Justin ;). (I still remembered I mistakenly refer Jason as Patrick once in the comments! :LOL:)
So to me, this is a show that I really won’t miss. Maybe Justin should bring Bob Walsh on as well 🙂
Thanks for having me on, guys!
I seem to recall saying a number of stupid things (dubious anyway) … feel free to call me on those … I’ll watch the comments and clarify here (or at least acknowledge my misstatements).
FOR EXAMPLE – my Azure comments were just off … after thinking about it, I’d summarize “what our customers want” as
1. Familiar experience for developing software that allows you to reuse what you already know from Windows Server, SQL Server, etc.
2. More “magical” experience for operating software – no need to manage guest OS or apply OS patches, easy to scale up and down, etc.
Several cloud platforms provide one of these two advantages. I’m not aware of any besides Azure that provides both.
So … that’s how I would have answered the Azure question after thinking about it for several days.
Thanks again for having me on your show. I look forward to listening to past episodes and learning more from you and your guests – as well as talking shop and becoming better podcasters together.
Cheers!
P
@Patrick
I was just researching which cloud provider to use for my SaaS business I’m developing right now and have a question regarding Azure. I hope you don’t mind answering it here since you talked about Azure in the show.
Can you maybe clarify for me what the difference between Azure Compute and Azure AppFabric is?
On this site http://www.microsoft.com/en-us/appfabric/azure/product.aspx it says under “Use Windows Azure AppFabric to:” #2 “Build custom enterprise and web applications in the cloud”.
But isn’t that what the web role in Azure compute is for?
thanks a lot.
Chris
No problem, Chris! If I can’t answer it effectively here, send me an email to Patrick.Foley@microsoft.com, and we’ll figure out a time to talk on the phone (or I’ll find someone who can state the answer better).
If you are going to use Azure for a Saas business, you are going to want compute at a minimum. The “web role” is where you write your ASP.NET code (including ASP.NET MVC, Silverlight via WCF RIA Services, etc.) to expose a web user interface to your customers. The “worker role” is where you implement background processes (can be .NET or native, including Java …). To communicate between those roles, use Windows Azure Storage queues. (See http://www.microsoft.com/windowsazure/storage/default.aspx)
For “familiar” relational database storage, use SQL Azure. For dirt cheap, apparently infinite, NoSQL-style data storage, use Windows Azure Storage tables. For files/blobs, use Windows Azure Storage blobs.
AppFabric helps in a couple of specific areas: first, the Access Control Service was just updated to make it easy to implement claims-based security – in practice, this should make it easy to let people use facebook, twitter, google, live, etc. to log into your site (ADFS makes it easy to integrate with enterprise systems, too). I have not worked through this myself, yet, but it’s high on my list of technical learning to-do’s. I’ve watched some of the videos about it on Channel 9, and it looks awesome. I’ll blog about it after I try it myself.
Second, AppFabric provides a “service bus” that allows you to create end points in the cloud that help SHIELD you from where your services are implemented. So if you have a service that you want to host on premise today but might want to move to the cloud tomorrow, you can create an end point using the service bus – when you move it, your customers don’t need to care. Yes, you still need Azure compute (web role) to actually implement the service. If you have no plans to move it, then there’s no need to put the service bus in front of it.
Those parts of Azure AppFabric are available TODAY, but there is a lot of near-term and long-term innovation happening in along these lines as well. Near-term, expect AppFabric to make it easier to implement caching, workflow (composite apps), and enterprise integration scenarios (think BizTalk).
Longer term … well, notice that Windows Server has features named “AppFabric” as well … think of the “magic” that will make it easier to scale and to move workloads from on premise to cloud and back again – all dynamically – those are the kinds of innovations that appear to be landing in the AppFabric brand, whether under Windows Server or Windows Azure.
Hope that helps … if not, please keep the conversation going here or reach out by email.
Cheers!
P
Patrick, thanks a lot for your detailed answer, it clarified everything I need to know right now.
Chris
Totally spot on with the “We have all done dodgy stuff” comments. As part of the “healing” process I will come clean here.
I used to write CAPTCHA decoders on request. I never asked what they were being used for but they where always targeting websites that offered sub-domains, so I am guessing spammers. It was mainly because I was looking for a practical application of my thesis and I just happened to fall into it. The money was pretty good for the work involved at about $150 an hour but I felt a little dirty at the end of it even though I totally believe that CAPTCHA’s are a waste of time.
I don’t do it anymore but that still doesn’t change the fact that I helped make the web a slightly worse place which I regret since I am a big fan of it.
@Ben Boyter – $150/hour?!?! Send me the link and sign me up 😉
@Justin That’s usually what it worked out to be. Most I already had code lying around for and just resold what I had already. Where it was from scratch it used to work out at about $90-120 an hour. I used to get repeat work fairly frequently since the owners of the sites would update the CAPTCHA just a little bit which required a tweak to the existing solution.
I got most of the business through an article I posted about a year ago? http://www.wausita.com/captcha/ It was my attempt to give back to people who were studying this sort of thing since there are no code samples on the web. It tends to rank pretty highly on searches for CAPTCHA’s.
“Who you are speaks so loudly I can’t hear what your saying” Benjamin Franklin.
That’s cool, Ben – out of curiosity, what do you think of http://research.microsoft.com/en-us/um/redmond/projects/asirra/ (I heard it described as “CaPETcha” – a name I like better). I wonder if the same skills that allowed you to break CAPTCHAs would also let you break this.
@Patrick That would be harder to defeat. The techniques I used were totally aimed at extracting text so I doubt they would even work in that case. I should out out that its also totally impossible for a vision impaired person to solve so not really that practical or even legal in quite a few countries now.
Another problem with those sort of CAPTCHAS is you can usually get the source data they had to create the images (since its based on some public photo-sharing site which tags the photos) and match against that.
I really dislike CAPTCHA’s with a passion. If you must have one, just have a “type human here” sort of thing which never changes. It will keep out 100% of spam bots unless you are being targeted in which case if there is enough reason to beat your CAPTCHA someone will do it.
*That should be “I should point out”
Good point about vision impaired – hadn’t thought of that.
My favorite part was that it required a live source of ever-updating dogs and cats – and allowed the user to adopt. I thought that was kind of cool.
Not a big fan of captchas, either. Not sure if we can live without ’em, though.
Thanks for your thoughts.