Don’t get me wrong we love the iPhone here at iTrinegy, it has loads of great features and you can actually read the attachments in emails!  But I have to say that in some areas it feels like work in progress…

Astonishing: no cut & paste, no search on emails.  Vital. Just look around the forums and you’ll see the comments and how people have tried to get around this.  It even seemed like Apple was originally saying we didn’t need these features!

Well, Apple have just announced V3.0 of the iPhone software and guess what, there they are: cut & paste, search (and not just emails either) and many more features like picture MMS and emails in landscape – oh yes I had forgotten about that one – great.   It’s arriving this summer apparently – yippee.

I haven’t delved deeply into the new features but it doesn’t seem like my other big gripe has been sorted – the ability to keep background applications running – particularly important for instant messengers which go up and down like the proverbial…  Yes, I know that you can currently jailbreak the iPhone to do this but for most people there are questions of warranty and legality surrounding that. And I don’t think that PUSH is the whole answer either. Why? – read on

Now we come on to battery life.  It’s really not very good at all.  I have to charge the iPhone every day or face embarrassingly running out of power.  One thing I’ve done to greatly improve this is to disable PUSH on emails instead going to a 15 minute fetch cycle.

Lastly a gripe on 3G Internet performance on the iPhone.  I don’t think this is an Apple issue though.  The number of times the iPhone has had 3G on its display but the performance of web page download has been atrocious.  If you read my previous post on the performance of mobile 3G you’ll have my view on 3G dongles.  Watch this space as we’ll be taking measurements on the iPhone and getting back to you with some numbers in a future post.

Recently, I got myself a 3G “dongle” so I could connect my laptop and be on the Internet anywhere at a “reasonable” speed.  If you haven’t re-visited these gadgets recently you might not have noticed that they’re up from the old 3G speeds of 384Kbps  to 7.2Mbps.  I thought you might like to know my experiences, and of course, being me, I took some measurements…

I went for Vodafone’s 3G USB Modem and Service  which is advertised as up to 7.2Mbps (download – subject to coverage) with a 3GB monthly data limit.  Most of the providers are advertising similar speeds. I went for Vodafone because they had coverage in the areas I wanted.

You get a very neat looking device that looks like a USB memory key, made by Huawei. When I got it home and plugged it in the software installed very easily on my Vista laptop from the USB modem itself, which was also behaving as a USB memory key and a CDROM.  So no extra CDROM to load – nice – well done Vodafone. 

It connected with 2 out of 5 bars (poor – it said) of signal strength and reported a 3G+ connection.  So, better than 384Kbps, but probably not going to be up to the full 7.2Mbps with that signal.

But, I noticed on first use that it was really “sluggish” (technical term) browsing Internet pages.  Sites were taking quite a while to initially connect etc and were slow to navigate around too.  On the other hand, after I navigated (slowly) to the BBC iPlayer web pages I was able to watch an episode of  Top Gear with no pausing to refill the buffer, so the speed was OK for this purpose.  This seemed like our old friends latency and loss at work – something never mentioned in the advertising blurb, which always concentrates on maximum download speed.  This is not Vodafone bashing by the way,  I think this is par for the course for 3G from any vendor, but it proves that the headline numbers which the advertising materials concentrates on aren’t the only important performance factors.

Didn’t have much time that day but generally felt disappointed at the sluggishness, and made up my mind to do some measurements another day.

————————————————————————————–

A friend of mine came round and asked if she could borrow the 3G dongle (for a couple of weeks) last Tuesday.  But I really wanted to do some measuring first, so I said she could have it on Wednesday evening as I needed to do some testing.

————————————————————————————-

7am Wednesday I started and was immediately “disappointed” – it didn’t feel at all sluggish, as it had the other day.  No matter – I think that shows the variability of Wireless Broadband. BTW, my laptop was in the same place give or take a few inches. Here are the numbers:

Signal strength 2 bars (out of 5) – Vodafone UK 3G+

ping google.com   (at 7am) – average time over 20 pings 136ms (Min 128ms, Max 147ms, 0% loss), so way-way more than the 31ms (ish) I get from my cable broadband. A quick tracert to google.com shows that most of this delay is in the first hop i.e. the wireless part.

I was interested in what speedtest.net would say about connection speeds but it was down, so I used Broadbandspeedchecker  instead.  It gave:  Download speed 1328Kbps Upload speed 348Kbp, so no 7.2Mbps, but reasonable.

I then downloaded a file from our web site: www.itrinegy.com.  The 14.3MB file (using http protocol i.e. a standard web site download) took 1 min 53 seconds to download – i.e. 1.012Mbps

————————————————————————————-
My friend came round on Wednesday evening (8:30pm) to collect the dongle but before installing on her laptop I thought I’d have another quick look at the performance on mine.  Sluggish! (Hooray).

“Sorry you’ll have to wait a few minutes – I want to get a few measurements”…

Here are the numbers on this occasion:

Signal strength 2 bars (out of 5) – Vodafone UK 3G+

ping google.com   (at 8:35pm)  ping times were all over the place – the largest value I saw was 1400ms!  Then it settled to between 500-600ms but shortly after was up to an average of 1000ms again.  These are satellite sized round trip times!

Speeds from Broadbandspeedchecker (for a like for like comparison): Download 824Kbps, Upload 323Kbps.   Speedtest.net was working again and gave Download 1510Kbps, Upload 346Kbps (Latency 473). This could have been a difference in when the test was performed, as I said the ping times and therefore probably other performance factors (note the latency of 346 measured by speedtest) were varying greatly. (Or it could have been the method used – I’ll look at that another time).

I repeated the file download (albeit a slightly smaller file) from our web site: www.itrinegy.com.  The 12.1MB file (using http protocol i.e. a standard web site download) took 2 minutes 35 seconds to download – i.e. 624Kbps

————————————————————————————-

So the second result was much more like the experience when I first plugged in the dongle.  Sluggish. But notice that the download speed for my file (624Kbps) and the Broadbandspeedchecker.co.uk download (824Kbps) and speedtest.net (1510Kbps) are not too shabby.

So why this sluggish experience when surfing?  Well, suppose that the latency (which represents a full roundtrip) to google.com averaged 1000ms through the whole test.  That’s a whole second!  If this is the first time we’ve visited a site (so it’s not in our cache) we’ll have to look up the name and get its IP Address – minimum time 1 second, then we have the “three way handshake” to establish the connection: 1.5 seconds  but the browser could do our page request (GET http://…) after 1 second.  Minimum turnaround for the first part of the web page another second.  So we’re at 3 seconds.  If the page needs a css file, graphics etc more seconds… 

So, the conclusion is – download speed here is irrelevant it’s all about latency (delay).

If you want to know your how your applications, web sites etc. might perform in mobile broadband (3G) you can re-create them, without the dongle, using our INE for Windows product.