3PHealth Blog

GoMo vs.. 3PMobile Performance Measurement Service – Head to Head Performance Test

Wednesday, November 2nd, 2011

Using GoMo (the new Mobile measurement tester from Google) I compared how long it took to download www.google.com vs. www.3pmobile.com

First the results for www.google.com – 0.241 seconds to load the Web page

Now the results for www.3pmobile.com – 1.294 seconds to load the Web page

Now for a reality check. What we need to determine is the size of the content. So using our own 3pmobile.com performance measurement service. Here’s where it gets interesting.

Gvs3P

In short Google’s Web page is 282k vs. ours at 55k. And yet it required a second more to download 80% less content. Hmmm.

A little too much Privacy due to “Do Not Track”

Monday, October 31st, 2011

This is interesting.

Google has a Web site that can format a Web page for Mobile viewing. You can find the it here – link

So I tried running www.sears.com through it. Here’s what happens when you have Do Not Track turned on in your browser.

And herein lies the “design flaw”. Cookies are there to enrich the Web experience. Turning them off kills the experience. This is what happens when Privacy is binary.

Privacy on the Internet is NOT “binary”

Monday, October 31st, 2011

I’m going to start this blog post by quoting the definition of privacy from Wikipedia. I’ve highlighted 4 key words.

Privacy (from Latin: privatus “separated from the rest, deprived of something, esp. office, participation in the government”, from privo “to deprive”) is the ability of an individual or group to seclude themselves or information about themselves and thereby reveal themselves selectively. The boundaries and content of what is considered private differ among cultures and individuals, but share basic common themes.

When something is private to a person, it usually means there is something within them that is considered inherently special or personally sensitive. The degree to which private information is exposed therefore depends on how the public will receive this information, which differs between places and over time. Privacy partially intersects security, including for instance the concepts of appropriate use, as well as protection, of information.

In summary:

Privacy is selective: I want to be able to share “selective” information
Privacy is personal: Everybody assigns a value to their personal information
Privacy is location & time sensitive: What happens in Vegas stays in Vegas
Privacy is valuable: My information needs to be encrypted and under my control

One of the current efforts underway on the Internet to increase privacy levels is something called the “Do Not Track” header (website). Here’s how it works:

You enable your browser (via a menu) to send a message to a Web server that says do not track me

Here’s what the Web server sees, a header:

HTTP_DNT=”1″

That’s it. Your privacy is “now binary”. 1 means don’t track me, & the absence of the header means track me. By the way there’s no support for this feature in any of the current Mobile browsers.

So how does this align with the Wikipedia definition. In short it doesn’t. By reducing your Privacy to a “black & white” (1 or a Zero) you lose any ability at selectivity. You remove the personal aspect, and device location is ignored.

Do Not Track is a start – but it’s far from complete. Any Privacy solution will HAVE to include the ability to control who sees what aspect of my information, and when and where they can see that.

That’s why when we built our 3P Mobile solution we included all of these attributes, and we built them directly into the browser. See more at 3PMobile.com To see screen shots of this running on Android and iPhone click here

The Real Mobile Performance Problem

Monday, October 3rd, 2011

This morning I ran a speed test using my iPhone tethered to my MacBook Air (via Bluetooth). I was not so much interested in the actual download speed (which I found to be pretty impressive using a 2 1/2 year old iPhone), but more in the latency numbers.

Here are the results – the top test is using my iMac to access http://speedtest.comcast.net Notice that the latency is sub 10 ms (great) and that my location used a server in Centennial CO (about 10 miles away). Overall I have no complaints at all on the desktop

The second test is the MacBookAir test via an iPhone. Note the latency – almost 1/2 a second. Also notice the server. Using my IP address it guessed that the closest test server was in AR.

Bottom line…

Latency (the time to locate something) is a big deal in Mobile. If you’re going to make a mobile page load really quickly you have to assume that there’s going to be a lot of system latency. Therefore you should focus your efforts on delivering the smallest amount of data (sub 150K) and ensure that it’s compressed and contains as little JavaScript as possible. (See my last post on “Sears shows what a really fast Mobile site looks like”)

APM – Identity Crisis or Something Else?

Sunday, October 2nd, 2011

Recently, I’ve been reading a lot about Gartner’s Magic Quadrant for Application Performance Monitoring, as each company mentioned is leveraging the coverage for it’s own promotional purposes.

GartnerAPM-MQ-2011

Having spoken with several of the companies on this list, and being the curious gal that I am, I thought I’d take a closer look at all of the APM Magic Quadrant contenders to see how each company positions and differentiates itself.

What I found was 12 unique variations on the “APM” theme from the 29 companies included:

IT Reliability™ software
End User Experience Monitoring
End User Experience Management
Application Performance Management
Application Management
End-user experience and performance management
Business transaction management
End-user experience and web application performance management
Real-time visibility
Unified service delivery management
Responsive process management
Application Performance Monitoring

Is this the inevitable product line extensions of a maturing network performance market? Perhaps it’s the natural fragmentation of an emerging market searching for its sweet spot? Regardless, I identified three overriding commonalities:

Each solution is very complex, providing lots and lots (and lots) of data
Most are desktop centric – there is little to no support for mobile apps or Web
Each company is primarily targeting IT or Network Operations

I did not see any targeting of the Business Analysts, Product Managers, Voice of the Customer executives or other app stakeholders. Hey, us non-techies like and/or need data, too. More than just network speed matters. Oh, and remember your employees and customers are accessing your customers’ sites and apps via mobile. The time to move is now. If an organization’s goal is to improve application/Web app performance, or any of the other 11 “APM” descriptions, then the real visionaries will be the first to fill in these blanks.