Was Sun right?
## My move to the cloud
It turns out that my move to the cloud might be what everyone else is going to be doing. Over the last week or so, I’ve changed the way I look at my computing devices. For a very long time (well just about forever), I’ve really only had one computer that I used as “the computer.” That’s despite having a ton of hardware laying around doing things in my basement and my desk. Those “other” computers were utility devices: my vmware farm for email/http/etc services, my mac mini for desktop/utility services, my netapp/open solaris boxes for file storage. My laptop was still the primary holder of what I considered critical functionality and data. If I lost or broke my laptop, I’d be in a world of hurt. Well, not really, backups are a good thing. I’d be in a world of “recover, waste time waiting, and then do work.”
With my acquisition of a truly powerful desktop (iMac 27″ core i7 – woot!), I needed to make a change. I’d prefer to work on my desktop when I can, and then go mobile when necessary – and do it seamlessly. The email part was easy, or should have been easy. I’ll post on that later. What was not easy was the data. In retrospect, it should have been easy, but I made it hard for myself. In my ultimate fantasy world, I would have liked a complete copy of all data, application state, and application configuration transferred from one machine to another. That way I could literally get up from my desk, move to the couch with the laptop, and just continue. Sure – I could have done that with remote desktop of some sort, but that’s not really an option when I’m on a plane or in a hotel with crappy internet. In the ideal world, I would only be sacrificing compute performance and screen real-estate for mobility. To get there, I played with a ton of sync options, both commercial and open source: rsync, goodsync, chronosync, etc. Unfortunately, none of them really give you the state of applications, and in the case of chronosync – your computers have to be physically close (as in the same network) to effectively keep them in sync.
My path to the cloud became clearer with the acquisition of the MBA 11″. Even though it’s a top of the line 128GB SSD model, it simply does not have the capacity to hold all of the data that I kept on my previous core machine – my 17″ MBP. That meant sacrifice. Out of sacrifice came clarity. Before this, I had not fully committed to the iMac being “the computer”. That’s because I wanted full access to everything while on the road. Well, the 11″ is going to be the on-the-road machine. I can’t have full access on it. The decision was simple: the iMac became the ‘puter. All of my iTunes and iPhoto stuff left the MBP and moved over to the iMac. With that move, I loose the ability to sync my i-devices on the road, but that’s ok. I’ve not been fanatical about that anyways.
All that was left was the problem of having my core important data available to me at all times on all computers. Enter Dropbox. Finally, I purchased a paid account on the dropbox service, and sync’d all of my core data to the cloud. My work flow had to be changed a little bit based on where I placed my stuff. Instead of ~/Documents/xxx, I now place it in ~/Dropbox/xxx. That service now automatically sync’s all data to the cloud and back to my devices – even my iDevices if needed.
### the network is not the computer
Sun’s vision was to make all services cloud based – including compute. The only thing you would need is an access terminal and your data and applications would bet there. The access device really needed only enough horsepower to run authentication, the network, and the display. VMware’s view and the rest of the VDI gang are headed down this same path. For much of the enterprise needs (think call centers and things like that), this is __the right way__. But – for me that means I have to be on the network. I’m not always on a reliable network or even a fast one. I have to have local compute and storage to do what I need to do. In all honesty, I think a very large segment of the non-home, non-drone corporate worker user base is in the same boat. The problem has been exactly the path that I went through: how to make the data and the compute always available.
### where I ended up at
After much mashing of teeth, and angst, I ended up here:
– The iMac is my central compute platform and also acts as the master sync for all data, including the iDevices
– Core data is hosted on Dropbox and automatically sync’d to all my devices, mobile or not (great value for $100/year)
– My MacBook’s have essentially become interchangeable. Use the Air for when I need light weight and simplicity (most of the time). Use the MBP when I’m traveling and need a desktop replacement (large screen, compute horsepower, etc.)
– The idevices (iPhone, iPad) have become more useful because I can use the data from Dropbox to do quick work on recent data
To accomplish this, I had to make one major workflow change: Close all apps at the end of the day. Because OSX is so reliable about sleep mode, I’ve gotten into the habit of just closing my laptop and moving on. Many times I don’t even save my work. Really. It is that good. Well, the sync thing requires that I do not do that. It’ll take a little while to break a 7 year old habit, but that can be done.
Anyways, if you think about what I have gotten to, it is this: My compute devices are interchangeable and I can select which one I use based on practical location requirements (i.e. am I sitting at my desk?, on a plane?, at a customer’s?, etc.). As a side effect of this, my data is now also safe. It’s on the cloud, and multiple devices. Loosing any device due to theft, negligence, failure, etc – means little other than replacement of the device. The important stuff, my work and data, are simply re-instated. Pretty damn cool.
Tags: cloud, sun, thin client, virtualization