Posted a reponse over on my blog, but figured I'd post it here as well in case you didn't find it (mostly because no one reads it. =))
--
Kent asked “Where’s my G-Drive?” where he pondered what it would be like to have your computer dropboxed (verbing the noun). He went on to dream of a time when your computer was essentially just a thin client connecting to a remote server run by Google, and a all your data was stored there. He listed off some really cool use cases, and then went on to explain why they weren’t a possibility yet. Bandwidth, privacy, space, etc.
But he’s thinking small. He’s also not ignoring some fundamental human concepts, and also why Dropbox is successful.
First, let’s look at why Dropbox is successful: automatic syncing. I can be on my computer, move to my iPhone, and still have all the files. Move to my laptop, and the files are there as well. You can even log onto their website into your account and access your files.
None of this is data backup. It’s all syncing. Once you realize why Dropbox is successful, you realize what you are actually looking for. You are looking for a device that intelligently syncs your data across all your devices without needing to sacrifice bandwidth, eliminating privacy concerns, and not requiring massive amounts of storage.
So, let’s move on, and discover what the service will be that takes the throne.
First, it will focus on syncing intelligently. This doesn’t mean syncing everything all the time, but syncing what you use. There are things my iPhone can’t open that I sync via Dropbox. It’s useless to download this. It’s also useless to sync this over the web when my phone and computer are on the same network. So, the idea is to sync, but only when needed. Sync-on-Read. Bandwidth here is an issue, sure. But only in speed, and frankly, when your living in an era when streaming HD movies is a reality and common place, it’s not a big concern. Let’s just say, Sync-on-Read is no more a problem then the current Sync-on-Save model currently in place. Also, once something is synced, it can be stored locally.
Next, since you only sync when you read, that means you need to connect to machine that holds the data. The downside here is that you need to connect to your home machine to access the data, and the machine needs to be powered on. But, I venture to guess that anyone using a system like this would keep their computers on anyways. After all, a computer shut off can’t sync regardless. Of course, you could also ensure that the system will figure out where it can get the file. If I have a file on multiple machines, and the main machine is off, the software can still grab the data.
The benefit to this is you still get to store you data locally. Local storage isn’t going away. People want their data. Sure, they want backups, but if their internet goes down, they don’t want to lose access to the data. And adding more data to the system is easy. Just add another hard drive like you do now. Also, the service provider need not store the data remotely. They just need to keep track of your various machines, and make sure machine A can speak with machine B. This eliminates privacy concerns. Instead of worrying how much disk space to rent, you just handle it yourself.
Essentially, you turn this from a push environment into a pull. You aren’t handed data, you ask for it. The service provider would manage things like figuring out which file is the latest copy, and keep track of local copies in case something changed and you didn’t want.
Kent’s original plan was “Install G-Drive, Tell G-Drive which files to sync, Wait 3 days for the magic to happen, That’s it.” In my plan, the idea can essentially be “Install G-Drive” and that’s it. Take your phone with you, and have access to the file. Technically speaking, it’s not incredibly difficult. It’s making the workflow easy. Dropbox easy.
Who will deliver this? Apple has the best opportunity to do something like this. They have the complete infrastructure, from hardware to software, to handle something like this. Controlling the entire pipeline, it’s really just a willingness on their part that is needed. Consider for a moment that they already have MobileMe which handles a lot of this. Microsoft can do this. Live Mesh and SkyDrive were initial attempts at this field, but still, it’s not a complete end to end solution like what I discussed. Google, if anything, is the one company farthest behind on this. They have various services, like Picasa, that already handle online storage. The problem is syncing everything together across the OS. MS and Apple both have their own operating systems and phone platforms they can use to bring it all together. Google is missing out (and while Android is awesome, it’s missing it’s older cousin, a Google OS that is actively being pushed).
Could Dropbox deliver? Maybe. At the software end, I think they could pivot fast enough and release a product like the one I discussed quickest. Will they? I don’t know. They’d have to charge for the service, and would someone want to pay for this? I would. A couple bucks a month to have all my computers synced like this easily would be nice.
I’d love to leave everything at home, and go, and not worry about forgetting to sync up the pictures this Christmas while I’m at my mother’s. If I want to show her videos, I shouldn’t have to plan that in advance.
Cloud storage is an excellent idea, don’t get me wrong. But I’d much rather have proper syncing first.
This entire idea of pull syncing is based on the premise that Netflix works, so bandwidth must be almost unlimited. But video streaming is a special case, in that the content is specifically engineered with streaming in mind, unlikely to be randomly accessed (but amenable to the procedure if it is), and streamed from a central server. Other content types often require the entire thing to be read before the data is useful. Also, most people's Internet connections have pretty lousy upstream (mine at home is about 14 KB/s, and at work it's about 200 KB/s).
Pull syncing wouldn't use any more bandwidth then "push" syncing. As I mentioned, the only downside is the upload speed for most ISPs is much worse then their download speeds. However, you have this same problem with push syncing. The only difference is that for push syncing, by the time you need it, it's most likely already been pushed up, and then down again.
I mentioned two things that solve for this:
1. Local network syncing. Much of my syncing occurs between my desktop and my laptop. Syncing shouldn't require an outbound connection (send up to dropbox, send down from dropbox).
2. Beyond large media files, most files people want to sync are still fast on high speed internet. Most pictures and documents aren't large enough to be a concern.
Finally, while most people's internet connections might be lousy, then they will suffer with the syncing regardless of the method used.
So yes, I do realize upload speeds are an issue, but it's not a big problem.
I think that "only difference" you mention is pretty crucial given the limitations we're talking about. If it's sync-on-save, then it's syncing while I'm doing other stuff. If it's sync-on-read, then it has to sync when I need the file, which can take a nontrivial amount of time. Looking at my desktop right now (which is just random stuff I happen to have received in the past couple of days — pictures grabbed off cameras, files emailed to me, etc.), I see that 10% of the individual files are over 10 MB and 25% are 1 MB or more. My browser's icon cache takes about 10 MB. Firefox itself is 53 MB. That's a heck of a launch time at 14 KB/s. A 1 MB file on my home computer would take more than a minute to open at work even if I use 100% of my home upstream.
> I think that "only difference" you mention is pretty crucial given the limitations we're talking about
Yes, it's a problem. A problem I talked about in the original post. However, it's temporary problem, and only for people that have to suffer with slow internet. For these people, even download speeds are horrible. Steam, for example, as wonderful as it is, is hated in places where download speeds suffer.
However, global internet speeds are increasing.
For me, I can stream my music and videos from my home to work. While I don't for a moment hide behind this problem, I won't pretend it's a road block.
Simply put: it's a problem as much as people without internet wouldn't be able to use it either. Enough people can use it that it would be useful.
"I venture to guess that anyone using a system like this would keep their computers on anyways" - way to limit your potential user base. I use dropbox to sync data between my work computer (turned on during the day, off otherwise) and my home computer (turned on in the evening, off otherwise) so the service you propose is useless to me and likely many other people who care about power usage.
That's all well and good, but meaningless. Who cares? A few other people that you do. My service is useless for people without an internet connection.
So, instead of just trying to point out problems, maybe you could point out a solution? As cool as dropbox is, it doesn't come close to solving the same problems.
I don't know. I think I'd take your post more seriously if it was something more than complaining and actually contributed some thought to the conversation.
I apologise for the unnecessary harshness of my previous post. The service you propose would probably be very useful for some people.
However the service would be fundamentally broken for people who turn off their devices or let them run out of battery or don't have an internet connection at the time they decide they want to access a particular file. Automatic syncing at the earliest possible opportunity (sync-on-save) is the best solution I can see.
> I apologise for the unnecessary harshness of my previous post.
And I apologize for mine. It was far too early, pre-coffee, and I should know better.
That being said, they aren't mutually exclusive services, however. And I did mention a push method as well. How this is accomplished could be as simple as a Dropbox method, maybe with a temporary store.
Being able to easily grab any file on one machine from another machine without having to do anything different (open find, browse to directory, open file and have it sync-on-read) and not being reliant on the disk space available on a remote service is key.
I'm not suggesting a dropbox like service would be useless. I can see the value in having an online store you can easily share with friends (something I wouldn't want my service as described handling, or at least, not in the default setting).
Consider the context of the discussion: Dropboxing your entire computer. That's silly. Rather, Dropboxing selected files, and syncing everything else on reading. Together they could work.
Edit: I realize now where the confusion might have arisen from, and why I think I answered it in my original post. I'm a moron. =)
> Sorry, ZumoCast is currently not supported on Linux
One of the big benefits of Dropbox (and something I forgot to mention) is that they work on a variety of platforms. However, I will take a look at that more when I get home on my Mac. Thanks!
Maybe I'm misunderstanding: but if you sync-on-read, the user still has to wait for the file to be downloaded if it's not synced up already. So maybe another model would be sync-on-save for those items that are appropriate for the device and that the user wants to have quick access to (whether because they've "favorited" them or via some other manual or automatic mechanism), coupled with sync-on-read for everything else?
Yeah, I forgot to mention forcing a push downstream.
But honestly, I don't thing that's important. Sync-on-Read is what you do now on the web. I click on a video on YouTube, and in seconds, I'm watching an HD video full screen.
I think the only hurdle is upload speeds offered by most ISP's is slower than download speeds. But even still, outside of HD videos, all the files I'd be interested in are quick downloads (pictures, documents, music).
Chrome can sync your bookmarks and extensions between browsers in different OS. So Google already has the technology albeit at smaller amounts of data.
Google claims that and yet when I used that feature across my work and personal computers I found that it often would duplicate my bookmarks or delete the bookmarks within folders, this was on XP and OS X. Back to XMarks and manual backups.
--
Kent asked “Where’s my G-Drive?” where he pondered what it would be like to have your computer dropboxed (verbing the noun). He went on to dream of a time when your computer was essentially just a thin client connecting to a remote server run by Google, and a all your data was stored there. He listed off some really cool use cases, and then went on to explain why they weren’t a possibility yet. Bandwidth, privacy, space, etc.
But he’s thinking small. He’s also not ignoring some fundamental human concepts, and also why Dropbox is successful.
First, let’s look at why Dropbox is successful: automatic syncing. I can be on my computer, move to my iPhone, and still have all the files. Move to my laptop, and the files are there as well. You can even log onto their website into your account and access your files.
None of this is data backup. It’s all syncing. Once you realize why Dropbox is successful, you realize what you are actually looking for. You are looking for a device that intelligently syncs your data across all your devices without needing to sacrifice bandwidth, eliminating privacy concerns, and not requiring massive amounts of storage.
So, let’s move on, and discover what the service will be that takes the throne.
First, it will focus on syncing intelligently. This doesn’t mean syncing everything all the time, but syncing what you use. There are things my iPhone can’t open that I sync via Dropbox. It’s useless to download this. It’s also useless to sync this over the web when my phone and computer are on the same network. So, the idea is to sync, but only when needed. Sync-on-Read. Bandwidth here is an issue, sure. But only in speed, and frankly, when your living in an era when streaming HD movies is a reality and common place, it’s not a big concern. Let’s just say, Sync-on-Read is no more a problem then the current Sync-on-Save model currently in place. Also, once something is synced, it can be stored locally.
Next, since you only sync when you read, that means you need to connect to machine that holds the data. The downside here is that you need to connect to your home machine to access the data, and the machine needs to be powered on. But, I venture to guess that anyone using a system like this would keep their computers on anyways. After all, a computer shut off can’t sync regardless. Of course, you could also ensure that the system will figure out where it can get the file. If I have a file on multiple machines, and the main machine is off, the software can still grab the data.
The benefit to this is you still get to store you data locally. Local storage isn’t going away. People want their data. Sure, they want backups, but if their internet goes down, they don’t want to lose access to the data. And adding more data to the system is easy. Just add another hard drive like you do now. Also, the service provider need not store the data remotely. They just need to keep track of your various machines, and make sure machine A can speak with machine B. This eliminates privacy concerns. Instead of worrying how much disk space to rent, you just handle it yourself.
Essentially, you turn this from a push environment into a pull. You aren’t handed data, you ask for it. The service provider would manage things like figuring out which file is the latest copy, and keep track of local copies in case something changed and you didn’t want.
Kent’s original plan was “Install G-Drive, Tell G-Drive which files to sync, Wait 3 days for the magic to happen, That’s it.” In my plan, the idea can essentially be “Install G-Drive” and that’s it. Take your phone with you, and have access to the file. Technically speaking, it’s not incredibly difficult. It’s making the workflow easy. Dropbox easy.
Who will deliver this? Apple has the best opportunity to do something like this. They have the complete infrastructure, from hardware to software, to handle something like this. Controlling the entire pipeline, it’s really just a willingness on their part that is needed. Consider for a moment that they already have MobileMe which handles a lot of this. Microsoft can do this. Live Mesh and SkyDrive were initial attempts at this field, but still, it’s not a complete end to end solution like what I discussed. Google, if anything, is the one company farthest behind on this. They have various services, like Picasa, that already handle online storage. The problem is syncing everything together across the OS. MS and Apple both have their own operating systems and phone platforms they can use to bring it all together. Google is missing out (and while Android is awesome, it’s missing it’s older cousin, a Google OS that is actively being pushed).
Could Dropbox deliver? Maybe. At the software end, I think they could pivot fast enough and release a product like the one I discussed quickest. Will they? I don’t know. They’d have to charge for the service, and would someone want to pay for this? I would. A couple bucks a month to have all my computers synced like this easily would be nice.
I’d love to leave everything at home, and go, and not worry about forgetting to sync up the pictures this Christmas while I’m at my mother’s. If I want to show her videos, I shouldn’t have to plan that in advance.
Cloud storage is an excellent idea, don’t get me wrong. But I’d much rather have proper syncing first.
--
http://blog.jasonlotito.com/uncategorized/whats-better-then-...