plightbo

[plightbo]

🎤 👋 This thing on?

[plightbo]

The Future of BMP

TL;DR — I’m leaving BMP but I hope the project continues under new leadership. I have some ideas about what work needs to happen and I’m here to help with the transition. Read more…

Over the last year I’ve come to two conclusions:

I am personally no longer able to commit the time necessary to maintain the BMP project, primarily because of Real Life: two young daughters (1.5yo an 3.5yo) and an increasingly-demanding day job.

The BMP codebase is somewhere between “middle-aged” (the SiteBricks / REST API layer) and “decrepit” (see history below).

Given these two facts, I’ve decided that the best thing for the project and the community is for me to (gradually) retire from the project.

I don’t want to leave the community in a lurch, so my hope is to recommend a technical direction (outlined here) and nominate a number of recent, active BMP contributors to carry the torch. 

I will, however, not be responding to any open pull requests or incoming support questions on the mailing list starting immediately. My commitment is that I support anyone else who wishes to continue this project or start up a similar-but-alternative one.

But first, a bit of history…

BMP (aka BrowserMob Proxy) dates way, way back… about ~10 years ago, when Selenium was first getting off the ground. You may not know, but Selenium wasn’t aways offered as a native “Web Driver” to various browsers. Back then, it was a simple JavaScript library that ran inside the web application.

Some of the committers (myself included) realized that this approach was limited and required that the application under test bundle Selenium inside of it. In order to bypass that requirement, we created an HTTP proxy injected the JavaScript in to each header of each page load and did all sorts of tricky stuff. Eventually we started doing really funky stuff, like doing man-in-the-middle “attacks” so that we could inject in to SSL. 

Meanwhile, I began to get in to performance monitoring and load testing and realized that the proxy could be modified to serve my desire to measure HTTP request performance. I ended up starting two separate startups and forking the Selenium code to serve my needs. While during my second startup (BrowserMob) I realized that it would be helpful if other people could get performance data from their Selenium tests, so I forked the Selenium code a third time and BrowserMob Proxy (BMP) was born.

I tell all this history partly because I’m a bit nostalgic as this is the last remnant of code I’ve dabbled with for a decade, but also partly because it helps explain why the project has the features and cruft that it has.

The Features aka The Good

Because of my participation and leadership in the web performance market, I learned a lot about the features people need when doing performance testing. Some of these features can only be done through an HTTP proxy. I won’t outline all of them (consult the docs or API if you want a full list), but the main ones I think matter are:

SSL support!

Control over DNS lookup behavior

Bandwidth simulation

Injecting headers in the request and response

Handling authentication

Backlisting individual URLs and URL patterns

And of course: extracting performance data in a useful format (HAR)

Any one of these features can be done using non-proxy technology, but it’s hard and requires multiple complex pieces of software to do it all. That’s why I believe that for general purpose usage, the approach BMP takes is a good one for most people. That said, I do discuss alternatives (proxy and non-proxy alike) in a moment.

The Cruft aka The Bad

There are two fundamental problems with BMP:

I’m a mediocre programmer :) Seriously, the code base is monolithic and tangled, making it harder to maintain than it should be.

It depends on a very old chunk of Jetty code that was created and donated by Jetty founder, Greg Wilkins, ages ago.

Issue #1 isn’t too big of an issue: we just need to get some better programmers involved, which has already been happening. 

Issue #2 is more challenging: in order to maintain SSL support and ensure that BMP works reliably with modern stuff like, say, WebSockets, we need to have a more modern proxy engine, but we’re unable to upgrade to something newer, including Jetty.

This is because the SslRelay code that does the man-in-the-middle termination is incredibly complex and hard to reproduce, especially with newer web servers like Jetty and Netty, that use async. 

The concept of the code is simple, but I will be the first to admit that I’ve never fully understood the deep, dark corners of the Jetty 5 codebase on which it depends and I’ve fortunately been able to simply tweak Greg’s code over the years.

Recommendation

My recommendation is that BMP needs a major overhaul — probably a complete rewrite. In particular, I believe that BMP could become a nice performance-oriented project that wraps around LitteProxy.

LittleProxy can do most/all of the basic proxy stuff BMP needs, including SSL, and it has most of the necessary callbacks/extensions to support the BMP features. It may be missing a few things, but I believe those can be addressed by contributing to LittleProxy and leaving BMP to focus on the clean, performance-oriented wrapper around it.

Writing it out it doesn’t sound like much, but in fact this is a considerable amount of work!

In addition to this core proxy work, I also think the project would benefit from pulling apart the core server from the client driver(s) for each programming language. On top of that, a nice web-based GUI could be incredibly useful for people just getting started or for people doing exploratory testing.

Alternatives

I believe that BMP is the best choice for most performance testing needs, but it would be prudent to call attention to some important alternatives.

On the proxy side of the house, there is Charles Proxy and Fiddler, which are both much more popular and well-known. The main downside for both of them is they don’t have all the features I outlined and they (IMO) are not quite as easily embedded inside of test automation systems.

More importantly, there are some interesting non-proxy solutions that should be considered, especially if you don’t need all the features of BMP and just want performance data.

If you’re using PhantomJS + GhostDriver, consider the getLog() API and ask for the type “har” and you’ll get back a very nice HAR file with all the timings of the assets.

If you’re using ChromeDriver, there is some really in-depth performance data available, though sadly not in the convenient HAR format.

If you’re using Firefox, with a little elbow grease you can get FireBug + the NetExport plugin to leave HAR files behind in a folder or sent to a beacon after every page load.

And lastly, many modern browsers now support the Resource Timing spec, which means you can make the JavaScript call of “window.performance.getEntries()” and get back individual timings for most assets on the page. Sadly, response code and file size are not included.

The New Guard

So who will carry the torch? I’m not sure! But the good news is that there have been a number of contributors and commercial organizations using BMP over the last few years. In particular, three individuals stand out: andyclark, jekh, and nite23.

I’ve already granted all three commit rights to the codebase and I hope that at least one of them steps up and carries forward the BMP vision. If they do, I’ll be here to help them along the way.

[plightbo]

BrowserMob Proxy 2.0 beta 9 released!

I can't believe it's been 6 months since the last release already! Though considering that I have a four month old infant (along with a 2 year old) it shouldn't surprise me that time seems to be flying by.

Anyway, lots of new stuff in this release:

Various improvements to make sure the HAR output matches the latest 1.2 spec

A bunch of improvements to the REST API (with docs!)

Replaced the User-Agent parser with one that won't take down the proxy if the online UA service is having trouble

Advanced feature to define which address to bind each individual proxy to

Bug fixes: concurrency problems, charset fixes, better logging, more tests, and more.

Note: in this release the request/response interceptor interfaces have been modified to take a second argument (Har). If you use the Java interfaces you'll need to update your code. If you use the JavaScript REST interface (introduced in the last release) you don't have to do anything.

As usual, you can find the latest download at http://bmp.lightbody.net.

[plightbo]

BrowserMob Proxy 2.0 beta 8 released!

It's been a few months and it was pretty obvious a new release of BrowserMob Proxy was in order. If you're impatient, you can get it here and skip this announcement. 

The most important thing to note is that I've forked the project back in to my personal Github repository. In addition, the website has been moved to bmp.lightbody.net. I did this simply because I no longer work at Neustar/Webmetrics and I'm the primary committer to this project, so I figured I should keep it where I know I'll have control over it.

Now to what's new:

A brand new test suite so I can finally make sure I'm not breaking anything too badly

Restored Java 6 compatibility, which broke in beta 7

Binary content can now be captured, resulting in base64 encoded content in the HAR body

We no longer modify the User-Agent header (it used to insert "BrowserMob RBU" in to it)

Post data (URL encoded and raw content) is now reliably captured in the HAR

This release also includes some REST API improvements:

Request/response interceptors can now be set via the REST API using a simple JavaScript scripting language (needs documentation still)

Automatic BASIC authentication can now be controlled via REST

Waiting for network activity to stop can now be accessed via REST

Network activity control, such as maxBitsPerSecond, is now accessible via REST

Control over connection timeouts is now available via REST

And more!

PS: Someday I'll get back in to more interesting blogging, but for now I'm just happy to post something :)

[plightbo]

Capturing errors in the end user experience

This year I again had the privilege of delivering a keynote during the Velocity Conference in Santa Clara, CA. Last year I shared my experiences, lessons learned, and cloud services I depended on during the lifetime of BrowserMob - from inception to acquisition (video here).

This year I spent my time on stage discussing the various types of tools DevOps people depend on to maximize performance, availability, and functionality. Stuff like HTTP synthetic monitoring, functional testing, application performance management (APM), browser-based synthetic monitoring, and more recently Real User Measurements. You can find the video here.

RUM was a big topic this year, especially because all the major browsers now support Navigation Timing. But I was surprised how much interest little sidebar on window.onerror got: I had quite a few people coming up to me after to learn more about JavaScript error logging.

Despite window.onerror having been around for years, very few people utilize it to gain insight in to the experience of their real users. For example, of the 60 sponsors at Velocity, only one (Dropbox) was actively capturing JS errors and sending the logs back to a central server.

What's interesting is that the equivalent thing for mobile apps has recently received quite a bit of attention, mostly due to several competing hot startups:

TestFlight

Flurry

Crashlytics

Crittercism

Atlassian's JIRA Mobile Connect

I never saw this kind of interest in JS error reporting platforms, and I think part of the reason is that people think of JS errors as "non critical" and not at the same level as an app crash. But with web apps as complex as they are today, a single JS error could be just as bad as an app crash, in terms of user experience.

There are few services out there that do JS error logging:

Fire Tower (recently acquired by Crashlytics)

ExceptionHub

Errorception

DamnIT

Most don't look super professional, with the exception of Fire Tower. And now that it's part of Crashlytics, it'll be interesting to see how the combination evolves and grows out the analytics around error reporting. I'm also curious to see if log management services like Splunk, Loggly, Sumo Logic, and Papertrail get in to the game of analyzing error messages and error rates.

[plightbo]

10 Tips For Achieving Email Bliss

Over the last few years I've learned that the best way to manage my inbox was to follow the Inbox Zero technique, widely advocated by Merlin Mann. As stated on the official website, it's not about literally having zero items in your inbox (though I do get there), but rather how much of your brainpower is dedicated towards managing your email.

Since I've been personally able to achieve a near empty inbox for the last few years, I figured I should share the tricks that work well for me. So here they are!

Tip #1 - Don't bother with automatic filters

To some this may sound absurd, but I actually believe that setting up advanced rules and filters actually perpetuates the email hell so many people live in.

It's natural to think that it helps: less stuff flows in to your inbox and instead gets sorted to appropriate folders. The problem is that you now have the constant pressure of checking those folders.

In my experience, anything I automatically filter in to folders or labels ultimately either gets ignored or just causes me to worry more about email, wondering what's hidden behind that folder - exactly what I'm trying to avoid!

For example, rather than subscribe to email delivery on some Google Groups mailing list, I almost always opt out of email delivery. Instead, I will search the archives and post my question. I'll then set up a reminder in a few hours or days to check back in via the web interface and see if my question is answered.

For any other distribution list that is important enough for me to be able to respond in real time, I let it flow directly to my inbox. I may still end up archiving or deleting 99% of the emails but if it's important enough to respond quickly, it's important enough to see in my inbox.

Tip #2 - Archive like crazy

With storage what it is today, there's no point in deleting. Even in archaic organizations that give you a puny amount of storage, you should be just shoveling email from Exchange to local storage. The primary benefit to archiving, or course, is that you can easily search for it should you need to pull up an old email that you glanced at and then tossed.

And that's the real point: don't spend a lot of time reading email. As long as you don't delete them, there's no real risk to just skimming and archiving. Obviously if it's an email you should respond to, don't archive it right away, but I bet at least 90% of the email you get can be skimmed and archived in a matter of seconds.

Tip #3 - Be concise

For the 10% of the emails you really do have to reply to, practice the art of keeping your replies concise. This is really hard to do - probably the hardest of these tips. In fact, at first you'll find that writing shorter emails actually takes longer, since brevity is such an artform. 

But it's also an important skill to learn, so keep practicing. Not only is it beneficial for your own time, it saves the time of those that you're sending email to. So if you can't respond in a few sentences, odds are that you should...

Tip #4 - Schedule meetings & reminders for the big stuff

Too many people (myself included, in the past) treat their inbox as some sort of todo list. But the problem with having your inbox serve multiple roles is it then creates stress and anxiety when you look at it. Is this email something I have to do? Is this just a new email? Should I reply now or wait until I do Task X? 

All these questions create uncertainty around every email in your inbox. It becomes less clear which emails demand immediate attention and which are associated with longer term tasks. Often the result is that you end up suffering in both directions: tasks that need to get done lag for weeks in your inbox, and emails that require more immediate attention pile up.

Fortunately, it's easy to avoid this: just make time for Real Work. That means if a request for something that takes an hour comes in by email, quickly reply by email saying "I'll have it ready by Friday at 3PM" and then block out 2PM to 3PM on Friday in your calendar to get it done. If there is no timeline associated with the task, putting it in your favorite todo list (ex: Flow or Toodo) is fine as well.

Tip #5 - Manage email in bursts

So far we know that we should archive the 90% of the emails that demand very little attention and that we should schedule meetings & reminders for the stuff that demands a lot of attention. But what about all the stuff in between?

That's where handling your email in bursts helps. For example, while I am always keeping an eye on my inbox throughout the day, I don't actually concern myself about it most of the time. Instead, I reserve most mornings and afternoons for about 30 minutes blasting through my inbox. It's at this time than most of my archiving, calendar'ing/todo'ing, and concise replying happens.

For the last week I tracked what my inbox looked like at the start and end of every ~30 minute email session:

You can see that at the end of every session I often have 10 or fewer total emails (this is for all my inboxes - personal and work) and no unread emails. I also, through practice, have learned how to send 30+ emails in a very short amount of time - even if some of them simply say "let's talk about this on Tuesday" or "I'll have this by Friday at 3PM".

Tip #6 - Wait a few days

Another thing to remember: even though the world seems like it's moving faster than ever, it's likely that very few emails that you receive really need a reply right away. In fact, even if you can respond right away, there is some value to taking your time before you send back a concise, thoughtful response: by delaying your responses you begin to condition your coworkers that they shouldn't use email like it's a realtime conversation tool (for you at least).

In fact, one really easy way to condition your peers is to... 

Tip #7 - Be shameless

Be shameless. Wait a few days and then respond and ask: "Do you still need help on this?". Not only does this help condition people that you will likely take your time responding to non-critical items, but it also sends a message that while you hope/expect that they've solved their problem, you're there to help in case they haven't.

In my experience, in a healthy team most email topics work themselves out before I really ever get around to them. It's still good to be kept in the loop and it's still good to ask if you need to get involved, but by letting issues sit around for a bit and then following up, you can force the right behavior, which is often communicating using realtime instead of with email...

Tip #8 - Get realtime: talk in person and use voice/IM

For some strange reason, millions of professionals have been groomed to send an email to ask a question, even when the recipient is sitting just a few feet away. Sometimes this makes sense, especially when the question requires a little more thought or can be answered asynchronously.

But most of the time people want answers now and still resort to email. Once you've conditioned your coworkers that they won't get prompt responses from you by email, start pushing them towards realtime communication channels like face-to-face conversations (shock!) and using tools like Skype.

Even better, if you're in an email thread that seems to be getting out of hand, avoid the temptation to keep replying (especially so when it's a frustrating/confrontational thread). Instead, offer to jump on the phone (RIGHT NOW if necessary) or schedule a meeting to work it out. A ten minute phone call is worth dozens of emails back and forth.

Tip #9 - Use Boomerang

Although you shouldn't be using email as your todo list, sometimes it's not practical to keep a todo item to track every little issue/request that you have of your coworkers, especially when you're sending out emails like "When will you have feature X complete?".

That's where a tool like Boomerang comes in. It lets you pick a time window for an email to pop back up in to your inbox if it hasn't been replied to yet. So you can ask Bob about "feature X" and tell Boomerang to remind you if you Bob hasn't responded in 3 days.

This is incredibly useful especially for mid-level managers (that's me!) who have a dozen or so team members and might have half a dozen outstanding questions for each one of them. None of them are important enough to warrant time on my calendar or an item in my todo list, but they are little ways I manage my team. I need to know if I ask a question and it goes unanswered, so Boomerang is a life saver.

Note: there are other competing solutions that work with other mail systems besides Exchange and Gmail, but I'm not aware of any better ones. If you know of some, please leave a note in the comments!

Tip #10 - Use conventions with your coworkers

Lastly, it can only help to use common conventions in your email subjects. For example: AR (Action Required), URGENT, and FYI are three common ones I use frequently. A simple way to encourage your teammates to use these conventions is to start using them yourself.

Over time your team should start picking up the conventions. But if that doesn't work automatically, don't be afraid to send out an email asking the team to agree on a few conventions and to understand that you won't review urgent items unless the subject says URGENT in it.

Final thoughts

It's important to understand that your inbox shouldn't control you. It's one of many tools that is there to help us do our jobs better. But it's up to us to decide when and how often we'll use each of these tools.

By following a few of these tricks and conditioning yourself and your coworkers to use email differently, you've got a fighting chance at experiencing the same email bliss I do almost every day: often less than 10 emails in all my mailboxes, and a day mostly spent outside of my email client.

Do you have your own email management tips? Please share them in the comments!

[plightbo]

Pretty cool video showing Steve Jobs and the original NeXT team in an offsite strategy session just as they are launching the company.

[plightbo]

From this year's GTAC: A good overview of Kevin Menard's vision for checking for visual consistency of look & feel of web pages across multiple browsers. Much of his lessons in this video are reflected in his startup, MogoTest.

[plightbo]

A touching from Mona Simpson.

[plightbo]

Where the heck has Patrick been?

A couple weeks ago, I started blogging again after a hiatus over over two years - my last post was in July 2009. I figured before I got too carried away with any more activity, it might be worth a quick recap of where the heck I've been!

Back in November 2008 I announced my second startup, BrowserMob. My goal was to re-invent the world of website load testing by utilizing real browsers (and Selenium) in the cloud. About a year later we launched our second product, a low-cost website monitoring service. 

During the summer of 2009 I did an exhaustive search for VC funding, getting several term sheets along the way. Just as an aside: it's amazing how much the power has shifted to founders in just 24 short months. Today it's relatively easy to raise a first round of funding on a $3M+ pre-money valuation. Doing so in 2009 was a different story!

But before I closed the financing, I was fortunate enough to have generated enough profits to simply hire two awesome engineers (and friends), Ian White and Rafael Ferreira.  We were able to continue to grow organically and profitably, racking up over 500 customers along the way.

And then out of the blue Neustar and a few other companies called to discuss an acquisition. I ultimately decided on selling the company to Neustar for many reasons, which I will write about soon. Needless to say, the experience and lessons learned along the way were life-changing. I met a bunch of fantastic investors, customers, potential acquirers, and founders of startups in the cloud space (ex: Twilio, CloudKick, AppFog, PuppetLabs, PagerDuty, etc). 

After the sale I moved from Portland, OR to San Diego to join Neustar and run the product and engineering teams under the Webmetrics product line, (which BrowserMob is part of). I've been having a blast since then - both professionally and personally.

Not only do I have a great team that I get to work, I have the backing of a publicly traded company with a $2B+ market cap that is 100% committed to expanding my vision of a simple-but-powerful suite of website performance testing products. And most importantly, since moving to San Diego my wife and I had our first child, a beautiful baby girl named Maggie Rose Lightbody.

Life is good and I'm glad to be stepping back in to the world of blogging, technology, and startups. Part of why I'm blogging again is that things have finally settled down. Between the chaos of the startup, the acquisition, the move, the new job, and the baby it was hard to do much else. But things are a little more calm now and I'm energized to engage with the startup community again and contribute to this blog much more regularly.

[plightbo]

Battlefield 3 looks awesome. I was a big fan of Battlefield: Bad Company 2. The entire BrowserMob crew - all three of us - played it a bunch. Doubtful I'll get to play BF3 too much with a 1 month old baby at home, but it'll probably steal every spare second I can find. Definitely hoping that the boring Call of Duty line finally gets eclipsed by Battlefield this holiday season!

[plightbo]

I love Steve and I love Apple. But it is annoying to see reports of Steve and others claim that Android ripped off the iPhone but few reports of the reverse. The new iOS 5 notifications are awesome. They also are a total copy of Android's. It goes both ways.

Read the original story here: From Steve Jobs' Biography: "I'm going to destroy Android" (shared from Google reader).

[plightbo]

Pretty cool Steve Jobs tribute video that walks through the history of the iPhone and all the technologies rolled up in to it (mobile phone, Walkman, internet, OS X, CPU, etc).

[plightbo]

Newsflash: people still buy software

I don't know the exact timing, but somewhere between 2000 and 2007, much of the Silicon Valley lost interest in the idea that a business could build software and that people would buy it. Sure, plenty of startups still went about doing the enterprise software thing - companies I worked for like Jive and Spoke Software.

But the excitement around those kinds of companies pretty much dried up after the bust and the valley re-emerged from it's hole a few years later. Valley 2.0 didn't seem interested in selling software. Instead, it was all about social media, advertisements, group buying, gaming (aka ScamVille), etc. 

The Independent Software Vendor (ISV) has been forced to stay exactly that: independent. Most investors you talked would tell you plainly that people don't buy software (or music, no one told Apple). They'd say that your idea isn't fundable - good luck to you.

I know: both my startups (in the software testing space) were extremely hard to get funding for, which ultimately turned out to be a blessing after I sold BrowserMob to Neustar and retained 100% of the equity. To be fair: My second time around (in 2009) I did begin to notice a difference in attitude and I met some great investors that encouraged me to sell pickaxes over mining for gold.

But something funny happened a few years ago. Maybe it was the launch of the App Store. Maybe the valley just matured a bit. Maybe it just took a few good consumer/prosumer/enterprise startups to break past the Facebook/MySpace/GroupOn hoopla. Whatever it was, the attitude started shifting and making software that people buy became (somewhat) cool again.

And good thing too. The big news this week is DropBox's whopping $250M round, $4B valuation, and their spurning of Steve Jobs. I'm one of their 50M users and I was just thinking the other day after having uploaded another 500MB of scanned files: "Wow, I bet DropBox knows almost exactly when I'm going to become a paying customer". And sure enough, they know it:

It’s only going to get better. That 96% of nonpaying customers is throwing their stuff into Dropbox at such a pace that thousands of people each day blow through the free 2 gigabytes of storage, and upgrade to 50 gigs for $10 a month or 100 gigs for $20. Even if Houston doesn’t sign up a single customer in 2012, his sales will double

Quick side note: DropBox, arguably the most successful example of a company that makes money the old fashioned way, applied for YCombinator in April 2007 (a good read, btw). I don't know Paul Graham, but I can't imagine many people funding the concept any earlier than that. For reference, by April 2006 (a year earlier) Facebook was all the rage and had already raised over $40M, and before that News Corp had paid nearly $600M for MySpace. 

And then there's Evernote, a service/tool that I'm already happily paying $45/year. I use it, along with my wonderful Fujitsu ScanSnap scanner, to turn my household 100% digital. As TechCrunch put it when reporting on their recent $50M, they want to be the antithesis of Zynga.

Setting aside all the various iPhone and iPad apps I've paid for (that's another post), I'm also happily paying (or have paid) for Pandora, Skype, Flow (before that Things), Reeder, 1Password, SmugMug, and more. So despite open source, despite the Facebookification of the valley, despite all the naysayers - all trends seem to be pointing to a renaissance of the ISV and the simple concept that PEOPLE WILL PAY FOR A QUALITY SERVICE (*gasp*).

Of course, it's not like the good old days - it's better! No more CDs or floppy disks. Even the desktop tools, like DropBox, 1Password, Flow, and Reeder, all have nice auto-update utilities so they are as easy to use as a website. But more importantly, both consumer and investor attitude seems to have shifted back towards the idea that it makes sense to pay for a quality service.

And that's what excites me most about doing another startup.

I know I'll always be in the business of selling software for greenbacks. My brain simply isn't wired to be successful at starting a social network, a game company, or a group buying site. While I don't expect to do another startup in the enterprise/IT space, I am super excited about building enterprise or prosumer applications and services in the coming years.

Knowing that there are 50M people supporting DropBox or 11M+ people supporting Evernote gives me the confidence that I too can build something of value and create a real, long-standing, successful company out of it. I may never know why things shifted back this way, but I'm certainly glad it has.

[plightbo]

We (BrowserMob) routinely launch thousands of instances on Amazon EC2 every day. Back in 2008 when I started the company, my bet was that as big as EC2 was then, it was only going to get bigger and cheaper. And boy did it! Today we're one of Amazon's largest users, in terms of peak concurrent instances/CPU cores. We don't quite do the 30K CPU cores that is referenced in this article, but we get pretty close. What's really cool is we do it all serf-service: our customers schedule load tests thatm under the covers, deploy massive amounts of hardware within minutes. It's really amazing to watch as we launch hundreds of thousands of servers every month. But as useful as load testing in the cloud is, it's still admittedly a simplistic use case that probably isn't changing the world as we know it. What I'm really excited about are all the people who now can rent supercomputers and do serious science. With Amazon's recent foray in to GPU compute clusters, I expect to see some really cool stuff come out of biotech and Amazon in the coming years.

Read the original story here: Amazon Lets You Spin Up A Supercomputer Cluster (shared from Google reader).

[plightbo]

Read the original story here: Predator-Inspired Ammo Backpack Cobbled Together By Soldiers In Afghanistan (shared from Google reader).

[plightbo]

People think focus means saying yes to the thing you’ve got to focus on. But that’s not what it means at all. It means saying no to the hundred other good ideas that there are. You have to pick carefully.

Steve Jobs