Craig Rood

As part of my homelab migration, I have a number of Ubuntu Server VMs to migrate out of VMware ESXi v8.0 into a Proxmox cluster running v9.1.

The following steps cover the migration process I have taken. These steps are likely to work against newer and older builds. Other Guest operating systems will vary.

Make sure you shutdown the VM in ESXi before importing, and of course, hold backups.

EXSi Storage Endpoint

First key step is to add the ESXi source as storage into the Proxmox Datacenter.

Navigate to Datacenter > Storage > Then Add storage. This will produce a menu that will give you a series of Sources, at the bottom ESXi will be available. Select this.

ID: identifier such as esxi
Server: IP Address or Hostname of the ESXi source
Username: Username of an account on ESXi that has required permissions – root will work, but not recommended.
Password: Password…
Nodes: Leave on all, this adds the storage to all nodes in a cluster. You can limit this to a single node if you have a specific machine for all imports.
Enable: Leave checked
Skip Certificate Verification: Checked

Backup retention is not required as this is not supported.

Import

Next step is navigate to the Node within the Proxmox cluster for which you want to import to. You should now see the esxi storage within the node.

Select the Virtual Guest you wish to import and press Import

You can adjust the VM before import, such as CPU and Memory. These can also be adjusted at a later date through the standard processes, so for the purposes of migrating over, I leave everything as default and press Import.

Starting VM

Once the VM has been imported, you can go ahead and startup the VM.

Use the console view to validate that the machine has been brought up correctly, and you are ready to make any minor adjustments that are required.

Network Device

It is typical that the network config within the VM will need to be adjusted as the host interface has changed.

Through the console on the running VM, login.

use ‘ip a‘ or ‘ip link show‘ to identify the device names, and status.

You can check for errors using ‘journalctl -u systemd-networkd‘

At this point it’s likely the main network device has a status of DOWN as the operating system is still configured to the old device from ESXi. This can be adjusted easily through the command line.

Navigate to the Netplan configuration folder ‘cd /etc/netplan‘

Use ‘ls -la‘ to list all config files in the folder

For myself, the config file is 00-installer-config.yaml, we need to edit this one using ‘sudo nano‘ – noting the file is owned by root. Entering the root password when required.

Using nano, update the ‘ethernets’ to the new device such as below. Static IP can also be configured at this point, but I prefer to assign via DHCP.

Once the file has been written out, run ‘sudo netplan apply‘ to apply the new configuration.

For best practice also reboot the VM – ‘sudo reboot‘

Once the VM has come back up – you can re-run the ‘ip a‘ command and verify the network device is now UP. As below.

From here your new VM will be migrated.

Last November (2023) Mark Gurman suggested that Apple could be working on their own cellular modems.

https://www.bloomberg.com/news/articles/2023-11-16/apple-project-to-replace-modems-made-by-qualcomm-falls-further-behind-to-2026

You can be forgiven for not really caring about this…. but I thought it was interesting.
I have a theory on why this could potentially be quite an important development – maybe more important than Apple Pay.

Back when Mark made his initial prediction, I made a small prediction that Apple will exploit this beyond just creating a chip.

I think Apple will 'backdoor' the eSim model and start offering data as part of the iCloud Packages. From that they can start working with the carriers to get the bits at prices they can't refuse. Things like roaming will be a thing of the past.

— Craig Rood (@CraigRood) November 21, 2023

eSim

To make this happen, Apple will utilise the eSim technology.

Apple has already moved into the eSim world with its iPhone and Watch range – instead of having a physical sim card in your device, you can load a data file (an oversimplified explanation) – this can act like a physical sim in conjunction with cellular chip in your device.

Apple currently have various different device configurations where eSim is available. These options allow upto 2 active eSims at once. with additional 6 saved on the device, meaning a single iPhone can have 8 in total.

Currently eSims are stored directly on a chip within the device, meaning if you would like to move your eSim to a new device you will need to get a new QR code.

Devices

Importantly, the devices won’t just be you iPhones and Apple Watches. These chips with eSim capabilities will be build into every, single, device. Including iPads, MacBooks, iMacs, and maybe even Apple TVs

iCloud

Apple’s iCloud service offers syncing and backup features between Apple devices using an Apple Account.

Day to day this is useful where you might have an iPad, iMac or even a MacBook for having a shared centralised photo library, music library, or even just messaging.

This will be relevant soon.

Apple Carrier

Apple wants a slice of everything, if you use Apple Pay they get a slice of the pie, buy apps or content? Apple also gets a slice. Listen to Apple Music? Apple gets a slice too. It’s only natural that Apple will be looking for ways to take a slice of the mobile service.

To do this, I predict Apple will be their own Mobile Virtual Network Operator (MVNO). This will allow them to look like a carrier without actually needing to build towers and physical infrastructure.

The physical infrastructure will be leased out and provided by one of the existing operators. These will then connect back to Apples datacentre where all the networking is done for the data.

Essentially, your provider will be Apple.

iCloud+

Back to iCloud…

If you are already an iCloud user, Apple probably already has your personal details, such as name and address along with card details. This is especially true if you are also an Apple Pay user.

Apple will incorporate the subscription for their data service directly into iCloud.

Personal Plan

This will mean that your primary phone, will automatically be assigned a data slot, and have access to the internet without a traditional carrier subscription.

I can envisage Apple also incorporating options for enabling this service for iPad and MacBooks. Quite simply, an additional service within iCloud can provide automatic eSim data service across any device with your iCloud username and password.

Family Plans

iCloud already has family plans built in where you can share storage allowance and purchases/services across the family.

It would be great if the service is also available to families through the sharing method. Meaning a single line can provide data service across the entire family.

Once we get this far, I would imagine Apple will have a limit on number of individuals per plan, and a data cap per family pool. With options to scale up and down on both slots and data allowances.

Example 4 Individuals with 500gb data per month, combined.

It would be nice to have unlimited data, but I can’t ever see this happening…. (also more on data allowances below…)

Roaming

One of the benefits of Apple running their own carrier, and likely across the world, is the ability to access other providers and get favourable deals across the globe.

Apple can pass these savings onto their customers, much more than what traditional providers can offer.

After all, as a network provider you wouldn’t turn down Apple. Otherwise your competitors will just say yes.

With this, roaming and traveling to other countries becomes so simple. There effectively is no roaming. Apple can provide a local sim automatically to your devices within the subscription.

Allowances & Costs

Noted above, It really would be nice if Apple had an unlimited plan, but that would then be potentially open for abuse across the subscription. A single iCloud plan an have multiple devices where a single sim can only service one device at a time.

To bridge the gap, I can see Apple offering a per GB/TB plan across all devices within the iCloud plan. Personal or Family.

Data is cheap across the world. Even at a resell, I can see around £10/$10 per 250gb. This might be more expensive than what you currently pay, but this would be valid across multiple devices, or family members.

Zero Rating

Not everything will incur a data allowance cost.

Across the traditional carriers, it has not been uncommon for services such as Facebook and Spotify to be zero rated.

Meaning Apples other staple subscriptions such as iCloud Drive, Apple Music, Photo Library, Apple TV etc. Will be zero-rated. Listening to music on your phone won’t eat into the allowance on your carrier plan. Sharing photos across the entire family won’t cost you anything.

Anything else..?

Well to be honest I could go on, I think the potential of Apple integrating a carrier service directly into their core products really is a game changer.

I could go further and talk about Apple TV or even a new Airport to provide internet access to all devices, not just Apple, Starlink…. but I will leave it here for now.

Certainly some big guesses in this quick post, lets see how many come true.

I’m fortunate to now be posting away using Gigabit fibre internet access – and its the future.

Prior to FTTP, I was on VDSL2 (FTTC) – Originally at the fastest speed of 80Mbps. Which actually is pretty good. It supports full HD/4K streams and with just myself using the connection done its job.

About 12 months FTTP, my 80Mbps became a 62ishMbps connection due to an ‘external fault’. For the most part, this wasn’t really noticed as 62 is also pretty acceptable, although frustrating knowing the service has degraded.

Dec 2019

Wasn’t really aware of it myself, but beginning of the year Openreach published plans to expand fibre into my local area.

Feb 2020

Openreach started to do some works locally from the exchange through the main roads. Again, at this point I didn’t think much of it. Heck, it’s certainly not unusual to have utility works around here.

Spring 2020

Later in the spring, I finally noticed. The telegraph pole that serves my house had a “Caution Overhead Fibre” sign added to the pole.

June 2020

Finally able to place an order for the new product. Selected the full 900/100Mbps package.

Initial install was due for 19th June, but with covid, this was pushed out to the 30th June.

Install was super simple, new overhead cable – about 20m.

Jobs done! Superfast internet access!

Few years ago an SDR (Software Defined Radio) USB dongle for £10. At the time I used it for simply picking up FM and DAB+ – for the price point it was ideal. I was aware of some of the SDR ‘features’ of this device but not until recently I decided to look into this much further.

SDR

Software Defined Radio, most commonly referred to as SDR is where the signal processing from a radio signal is done in software rather than hardware. This allows for signals that couldn’t otherwise be decoded – such as ADS-B (simply information planes broadcast to let others know where it is etc), be done by simply installing software/plugin.

Device

Most popular hobby SDR devices are based on the RTL2832U chipset. This chipset was fairly groundbreaking at the time, a small and very cheap USB device was capable of doing exactly the same as a device that costs 10x. Very quickly the software library for these small USB devices grew and 100s of packages and plugins are available for hobbyists and professionals alike. Even today new devices are being produced based on this chipset with SDR in mind, rather than simply to be a TV/Radio receiver.

I’m not going to make any specific recommendation around what SDR device to get. There are plenty of reviews online including YouTube which covers all price ranges from £10 and above.

Setup

To run the software you are going to need a machine that hosts the device. More recently Raspberry Pi hosts are become more popular. This is primary due to Raspberry Pi having fixed hardware, so images can be produced and deployed to give a plug-n-play experience. The downside to this is, unless you have some experience with Linux, it can be difficult to customise. Preferably, a windows host will do the job, which I will be referring to in this post.

Installation

USB Receiver

Probably the hardest step is getting the USB Receiver installed onto a Windows. This is especially true with driver signing and compatibility. Zadig will install the driver for you, for beginners, downloading the community SDR# package is all you need. This includes the required drivers and software to get started.

This may require a few restarts of you machine, you may also need to remove the drivers and re-install for Windows to finally pair with the device.

Antenna

Your receiver should already come with an antenna. These certainly have limited mileage. Again, I’m not going to make specific recommendations into what antennas to purchase, but its important to note that both the antenna and positioning will change what you can and can’t receive.

Software

Now to the good bit, to start tracking flights, you will need 2 additional bits of software (also free).

Firstly, dump1090-win – (direct download link). Simply download, unzip and run dump1090.bat. This will automatically connect to your USB device – so make sure nothing else is running that may get in the way!

If successful, you should see a window that looks like this:-

Above shows information about the planes in your area, such as flight number, altitude, speed and location. Whilst this is neat, it doesn’t mean much visually. This is where our second piece of software comes in.

This one is called Virtual Radar Server. This takes the information shown above, into an actual map and visualise the planes in real-time.

Download the latest version. This will install onto your machine, unlike dump1090.

Once installed and running, setting up the receiver to hookup to the already running dump1090 is quite simple.

Go to Tools > Options. This will present the settings screen with a sidebar on the left. Select Receiver > Receiver. Click Wizard button on the top right.

Choose Software Defined Radio, then Dump1090. Once complete you can go back to your main Virtual Radar Server screen, as shown below.

By clicking the blue link near the middle of the page, a web browser tab will open up giving a visual on a map all the planes the SDR device can pickup.

There are quite a few options to play around within the Menu.

Limitations

Before I close this one off, it’s important to understand that both cheap devices and antennas have limitations. Personally most planes that I can pickup must be within around 20-30mi radius. Also, they also must be above 10,000ft. As ADS-B works from line of sight, having a small antenna inside is going to severely limit the reception.

More…

ADS-B is a great source of data to be use for other projects and learning. both pieces of software above are free to use, and can easily be interrogated with your own creations.

SonicGoose uses the publicly available data from himself to track flights on his website, as an example. He also has some tutorials and other software which may be of interest.

Further Reading

RTL-SDR: Seven Years Later

Realtek RTL SDR £15 Dongle Explored

RTL-SDR Tutorial: Cheap ADS-B Aircraft RADAR

One of my local dev environments run WinNMP – a Windows development stack package with Nginx, MariaDB, PHP etc. Great little tool to hack with.

Over the course of around 3 years my database has grown to around 300-400MB. In itself its relatively small for a database, but for test data is quite chunky.

Now, more recently MariaDB has started to play up, when running WinNMP often MariaDB would enter a crash recovery cycle, never actually starting up properly. mysql.log would read as below, with not much more to explain what is going on.

2019-12-05 17:59:26 0 [Note] Recovering after a crash using wt-nmp-bin
2019-12-05 17:59:26 0 [Note] Starting crash recovery…
2019-12-05 17:59:26 0 [Note] Crash recovery finished.

PHP would also throw errors

[05-Dec-2019 17:59:34 UTC] PHP Warning: mysqli_connect(): MySQL server has gone away in index.php on line 10
[05-Dec-2019 17:59:34 UTC] PHP Warning: mysqli_connect(): Error while reading greeting packet. PID=7336 in index.php on line 10
[05-Dec-2019 17:59:34 UTC] PHP Warning: mysqli_connect(): (HY000/2006): MySQL server has gone away in index.php on line 10
[05-Dec-2019 17:59:34 UTC] PHP Warning: mysqli_error() expects parameter 1 to be mysqli, boolean given in index.php on line 10

These errors simply make effort to describe what is going on. After plenty of Googling, and even some desperate Bing searches. I came across an almost decade old post that outlined an issue around replication. This unfortunately is no longer available, which is why i decided to write this one up – least not for my own reference.

Turns out that WinNMP – formerly WT-NMP runs replication locally, and some action is causing this to break between the master and slave.

Luckily, with WinNMP being portable, and the database being stored neatly within its own folder, I was able to make plenty of backups and play around with recovery with little chance of losing any actual data.

To restore service to MariaDB, firstly I kill any running applications related to the package. Backup the data directory. Remove multi-master.info, all files beginning with relay-log*, then finally all files that have *-relay-bin*

Simply spin the server back up and the MariaDB/MySQL works again.

I still continue to have the same issue about once a month, so its not likely a permanent fix. So I would certainly recommend rebuilding the database server at this point.

I originally started to pencil in this post much earlier in the year, but it got left in the drafts folder for quite a while, but I thought it was time to get this one out there with the very recent launch of Disney+ over in the US.

This year alone, here in the UK we have seen the launch of a Sports news subscription service called The Athletic. Whist It focuses on Football for the domestic audience, the service also follows most of the major US sports (including NFL and NBA).

In the past few months we have seen the announcement of even more subscription model based services. Most notably, Disney+ and the UK launch of BritBox.

Now traditionally if we talk about subscriptions you may think of magazines or newspaper subscriptions, beyond that, we didn’t really have many other products.

With a subscription in the physical world, you are guaranteeing the publisher a set monthly/weekly fee for a discount in their product. As they are delivering a physical product, a portion of the subscription goes directly to the printing and delivering.

As we moved into the digital world your typical subscription is no longer physically delivered, you are being provided a digital service.

The cost of providing a service to a customer is not only lower overall, but if the product is not full consumed, it provides ‘free’ income.

Mobile Phones

Good example of this model and approach can be seen by looking at how mobile phone contracts have evolved over the past 10-15 years.

Originally you had to pay for line rental, then pay a set price per minute or text message. After a short while the providers realised they can start bundling up the product and provide both minutes and texts into the line rental. This was great for the consumer as they had a known monthly price for their typical consumption rather than receiving a different price each month depending on how much the phone was used.

Bundles where selected to meet the customers requirements – where a customer typical uses of 200 mins a month they might select a bundle of 300 mins. The provider now gets a fixed income from the customer, with the knowledge that each minute consumed it paid for, and anything left off is again, ‘free’ income.

As more customers use the same package, and the costs to serve the customer decreases, the provider has a higher profit margin to work with. As a typical customer consumes within their package, the ability for the provider to ‘oversell’ increases.

If 10 customers existed on the 300 minute package, and only consumed 200 on average, total consumption is 2,000 minutes where 3,000 have been paid for and costed.

As we moved through the years up to today, most providers will now give you unlimited minutes. Whilst partly this is due consumers making less calls, the cost to provide the calls has vastly decreased, the provider has the ability to leverage the £20/£30/£40 bundles. The cost to provide the service is low, and the actual costs of the customer usage very low.

Similar model was also seen with mobile data, today GiffGaff can offer 80GB for only £20 a month. Whilst providing 80GB of data absolutely costs more than £20 the economics across the entire customer base means they working on a decent margin through the ‘oversell’.

Cloud Storage

Moving further into the digital world, the model of overselling can be seen more prominently.

Back when Gmail first launched they provided a whopping 1GB of storage for EACH address. At the time it was noted that each GB would cost Google around $2. The assumption also being that each address would never use anywhere near that amount of data – attachment limits also applied.

Now 15 years later it costs nowhere near as much as $2 per GB, £1.59 a month gets you 100gb of storage across the entire suite of Google products. Again given most customers will never use the full allocation, Google can aggressively price their bundles and quotas to maximise revenue.

Microsoft 365, SaaS

For Microsoft 365 the pricing model and principles outlined above clearly demonstrate the guaranteed income and potential margins for selling added service such as storage. With Microsoft 365, they are also importantly delivering an application, which has fixed development costs.

This was actually an easy decision for Microsoft as they already have the suite of applications, and an active roadmap for both home and business customers. To build out Office into the 365 package. They can not only guarantee a income each month – rather than sporadic retail purchase, they have the ability to up-sell and ‘lock in’ the customer into the future.

SaaS (or Software as a Service) is forever going to grown in the next decade as the small monthly income provides a much more stable income for development whilst also providing continuous product updates at the ultimate cost of lock in. The consumers income is gradually going to get eaten away by services and products that are not actually need, and certainly don’t provide overall value.

By now you have probably already heard about Bitcoin – an Open Source peer-to-peer digital asset where no single entity or individual has control.

Whilst its likely that the lack of control will be seen as a play against central banks, or even governments – the most interesting aspect is the peer-to-peer networking which makes Bitcoin permission-less.

January 2019 marks the 10-year anniversary of the ‘Genesis’ Block being created (mined).

In the same way the first iPhone changed the whole concept of mobile communications, the moment the very first block was mined will be remembered as one of the most important moments in money.

Whilst Bitcoin in itself may never be the token, or currency that we will all use day to day. It’s inception has given us a small taste in , and the principal of secure, peer-to-peer money.

But what does all this really mean?

Money is typically defined as a token used for the means of exchange for goods and services. Whilst money is more often than not spend digitally, this is a represenation of the phyisical token – cash. Both co-exist within the same financial system to makeup what is our economies.

Bitcoin is an entire digital token, its impossible to exist as a physical entity due to the nature of the cryptography which secures the token.

The magic isn’t in the idea of an entirely digital token – these are nothing new. What’s different is how the token has no single point of control.

The lack of control is ultimatly that which gives Bitcoin it’s value. All other digital tokens, or even traditional money are controlled, and effectively has it’s value already defined. Bitcoin moves away from this concept and has a value of that defined by the users of the tokens.

Bitcoin has the potental to become the global digital currency.

We have just come out of 2018, and I would consider it to be one of the slowest years yet for innovation in tech.

Not say that their hasn’t been any innovation or important events that have taken place – I see examples such as Lightning Network launching, Dropbox & Spotify IPOs as pretty significant. The three examples sit on the back of innovation from almost 10 years.

If we look at other innovations from around the same time (2008) – we see the largest social networks of 2018, finding their feet and really taking off. Facebook, YouTube, WhatsApp, Instagram, Twitter, Tumblr.

Even more importantly for what the future of TV and digital entertainment – Hulu was not even a year old, Netflix was itself only a year old at that point and Amazon Video was not far behind.

It would be wrong not to at least mention the launch of the original iPhone which came in 2007 – and change everything.

So my thought is; where is the innovation now, and what will come next?

Decentralised Internet

One area I am certainly following is the push to decentralised the internet – and embrace the ‘original’ winning formula (the competing ‘internet’ was centralised – and quite unlucky to not win out).

This is actually one of the reasons I have started blogging again.

We spend so much time in these walled gardens, creating a new version of ourselves to simply gain satisfaction from our friends and acquaintances – or even strangers. Not only is it unhealthy, we are relying on algorithms to decide what information we consume.

Decentralised is having your content spread across the internet in various entities brought together by you device. From this point independently built apps – (how we have many different web browsers to view the same websites) can decide for itself how it shows you the content, and what feature are important to you – (again how you may choose browser extensions)

Having a single service – such as Facebook, hold not only all your likes and interests, but your reviews, location history, family photos, contacts is only a privacy disaster waiting to happen.

Having for example photos stored on a separate service to Facebook or Instagram could provide better quality image storage for a small fee – something that can’t be bought from Facebook. Whilst products such as Flickr and 500px do have higher quality. They are still closed platforms, and couldn’t be inserted into a social feed.

Being open allows for entities to compete on feature lists, interchangeability and pricing.

Now, the decentralised internet can’t operate for free – without yourselves becoming the product through targeted advertising. Which is how all the free services are ‘paid’ for. Open creates a competitive market that will in no doubt undercut the existing paid services, and have more comprehensive feature sets compared to the existing centralised platforms.

Self-Driving Cars

Forget the dream of flying cars, there are too many restrictions, big issues around noise, and very limited infrastructure. The future is self-driving cars, and it’s here already.

Waymo One, launched their self-driving service and is available to the public in small area of Arizona. Won’t be long before this expands and be more readily available.

Uber are already trying to figure out how they fit into the self-driving world, and personally I think they will follow the airline model and simply sell seats. Third party companies will do the heavy lifting in terms of managing fleets.

Almost all car manufactures are working on self-driving vehicles, along side all electric, this will push them into the next full cycle of vehicles. Each and ever car currently on the road being replaced with electric cars – with varying levels of autonomy. Primarily due to them being cheaper to run, less moving parts, so less repairs, and longer life parts including batteries.

From the first cycle, we will then see another cycle where the autonomy will be far superior than the first cycle, safety will be proven and the insurance industry will price human drivers out of the market. Batteries will be even better, motors will be even more efficient , and most noticeably, car designs will massively change.

My prediction is 5 years at most for self driving to be common enough that more than 50% of the population has used one at some point in their life. Likely with a ‘driver’.

First cycle would also start around this time, with the second cycle hitting at most, 10 years later.

This is potentially huge business for the car industry, and where many would predict the end for car markers, this is potentially just the beginning.

Just finished watching the Google IO 2012 Android Keynote, and It’s made me wonder, Does Apple have a heart?

Apple are very notorious in how they conduct their keynote events, so much so that Google, Microsoft and the rest try and emulate Apple in presentation. Apple are very slick every single word is carefully selected to be informative and portray the correct information – same came really be said about Google today. Whilst there was a clear script, there could have been a better structure.

Google’s presentation was all about people, and how their products work in the real world. So much so that genuine photos from Google employees where used, from family photos to a wedding ceremony. Apple on the other hand do try and create the ‘people’ feeling, but fail pretty miserably.

Apples demo reel is about showing their users how they should use Apple products and telling them why they want to use them. From the iPad reel with the product being used in a hospital to a Facetime video call with a ‘family’. It’s all a Hollywood production, and an clean and boring one at that.

Don’t get me wrong, I prefer Apple’s way of working, I lost count of how many times I got ‘bored’ watching the Google keynote just an hour or so ago. Google reach out and do something important though, they don’t tell you what you want to see – they let you decided what you want and allow you to become attached.

This is at the core of the companies, Apple is all about selling you a fancy new computer that you think would be perfect for you because you own a fancy camera, but its not what you NEED, and often because of the price tag It’s seen as an important investment.

Google present you a online documents editor, they don’t tell you you want this because It’s better than everything else, they create a product because It’s what they feel people would want, present it, and allow the masses to decide if they even want to use it.

It may seem like a silly comparison, but again, at their cores the important part is that you really do want to use their products because you like them emotionally, not because you have already invested.

I love my iPhone and Mac, not because of the investment, because they are good. My fear is though, that there’s not connection between me and Apple. Tomorrow I would be more than happy to replace my Mac with a Windows 7 Tower and pick up an Android phone to replace the iPhone. After all, the functionality is still the same. Even if the methods are slightly different.

So maybe after all It’s just me, but I present to you the question, Does Apple have a heart?

Apples iMessage system, officially released last October along side of iOS5, allows iOS users to communicate between devices. Sounds fantastic, and for the most part it is. Least not for the feature which means SMS messages go over iMessage where possible.

Yesterday (16th Feb) Apple released a beta of Messages for Lion (Mac OS X 10.7) users, a core feature of Mountain Lion (OS X 10.8). Messages give OS X users the access to the iMessage system giving communication between OS X, iPad, iPod and iPhone in any combination.

iMessage works in a similar way to a SMS or IM message in that messages are sent to a iMessage address, be it a phone number or a verified email address. The sender of the message also sends their Caller ID to the recipient to create a communication stream. Meaning that the same sending location receives the returning messages.

Problems occur when devices such as iPad, iPod and OS X cannot receive phone number iMessages, meaning that when messages go out from such devices a different Caller ID is give. This creates another stream resulting in 2 communication streams between a single pair of individuals. Potentially very confusing. The target of these 2 streams become indistinguishable. How does the end user know which way the messages are being sent? Messages targeted for the phone may be sent to the desktop or iPad resulting in these messages being potentially missed for long periods of time.

Solutions.
iMessage needs a status indicator and a device logo to help users identify which streams would work, falling back to SMS when theres no other option.
iMessage addresses need to be consolidated to a single contact card.
Possibly a single stream between individuals rather than addresses, with a switch icon to identify where the message is being received – gets confusing though.

Finally, allow none iPhone devices to receive phone number iMessages.

I can only assume Apple have seen these problems coming. I do hope however that It’s not a solution as part of iOS6 – leaving many iMessages out.