This is the last battle for Google. Enable the interaction between your browser and the applications installed on your Mac. Think of a seamless integration between web and local filesystem.
This is what Google Drive can do if you use it via Chrome. The ability to launch locally installed applications from the browser is immensely useful for Google Drive users. You can now store in Google Drive obscure file types knowing that you’ll be able to open them with your favorite software.
The example that Google shows is about storing Photoshop images in Google Drive. By simply hovering over the image you can choose to open those images in the Adobe software.
With this new functionality, Google Drive increases its usefulness tenfold. And if you are in the market for a , Google is likely the way to go.
Other services that I have taken into consideration but not used:
Amazon Drive
You may have noticed that I haven’t included Amazon Drive in this speed test. To my surprise I discovered that anymore:
The Cloud Drive desktop application is no longer available for download from the Amazon.com website. If you currently have the application installed on your computer, however, it will continue to upload files, photos, and videos to your Cloud Drive account.
Wuala
I meant to test Wuala too. However, when I installed the application, I was prompted to install the Java Runtime Environment as well. In 2014, the need to install further software in order to use an application is annoying, so I decided not to test this sync service and deleted the application straight away.
Copy.com
When registering for a free trial, I was unable to confirm the email address with the service. I tried doing it for two days in a row, and on the third I simply deleted the application from my Mac. Too bad.
Warning
I acknowledge that this testing methodology is far from being scientific.
My setup and test methods might have affected the results. Other people might even get different values, maybe opposite from what I am sharing in this post.
The results posted on this blog are for personal use only and do not constitute a definitive proof on what service is faster.
More tests, using different scenarios, different networks, and a systematic way to time events are needed to calculate average times, variance values and to draw some more informative conclusions.
In any case I hope you’ll find this post useful, happy reading.
Methodology
I have simply installed the Dropbox, Google Drive, OneDrive, Box, SpiderOak and Syncplicity desktop applications on my Retina MacBook Pro 15″ and started uploading/deleting files.
I’ve used two different files for my tests.
AAC file
The song was The Miracle (Of Joey Ramone) by U2. Its size is 9.3MB. With this file I carried out two sets of tests:
An upload operation through the Finder by a simple copy/paste command. I did this three times for each service, on different days and at different times of the day to get more real world data.
A delete operation of the file from the services’ web interface. By executing this test, I wanted to measure how responsive the cloud service was to events triggered in the backend.
Dropbox and SpiderOak seem to be employing similar techniques to speed up multiple uploads of the same file. That is, the system checks the file’s hash and if it’s already present on the server then it simply retrieves that file from the backend instead of uploading it again. The result is that successive uploads of the same file are faster.
It also means that tests like mine fail because the second upload of the same file is always faster than the first one. Before each upload — in an effort to deceive the backend server — I edited the ID3 tags of the music file. That made the server think it was uploading a new file altogether, helping with the results of my test.
Directory with a complex folder/sub-folder structure
For this test I chose the CloudKit folder retrieved from Users –> [user] –> Library –> Caches. In my case this folder was 56.7MB in size.
The reason for using such a folder is because I wanted to assess the behaviour of all these cloud services when you try to sync a file with a deep folder/sub-folder structure.
In fact, in one of my previous tests Google Drive failed miserably when trying to sync these type of files.
Results
You invariably introduce errors when executing tests in this primitive way. For this reason I chose not to consider fractions of a second in my measurements. All data you see is rounded to the nearest second.
To minimize the effect of other applications accessing the Internet, before running my tests I also quit all browsers, Mail.app, RSS reader, Spotify, and stopped Time Machine.
– Broadband connection with 120 Mbit/s down, 10 Mbit/s up
AAC file upload speed
In this test the surprise was iCloud Drive and Syncplicity. Both services were on average the fastest to upload the 9.3MB aac file.
OneDrive’s results are bad because of the time it takes the applications to initiate the sync. In my tests it looks like the polling time for OneDrive is greater than 10 sec, which negatively impacts the overall results.
SpiderOak’s results are the worst of the bunch due to the client side encryption that this sync service carries out before uploading the file. In this case you cannot really blame SpiderOak. It all comes down to compromises: Security vs. speed.
Folder upload speed
I carried out this test only once for each service. Statistics theory says that you cannot draw any conclusion with one test only.
The reason for executing a test with a complex folder/sub-folder structure was to confirm some doubts I had after running a test a couple of years ago.
More specifically I believed that not all sync services work well with complex directory structures. The results above confirm my suspicion.
The fact that iCloud Drive is the fastest is probably a consequence of Apple optimizing this type of data transfer. All iCloud Drive data is stored as a very complex directory structure, so Apple was forced to optimize its service for it.
Google Drive on the other hand, shows the same problems I saw some time ago. It is unable to quickly upload folders with non-doc type of files and arranged in a complex way. It is obvious that Google doesn’t have any interest in optimizing this aspect of the data transfer, preferring to put its efforts on the upload of more common doc type files.
Delete speed
The delete speed test aimed at measuring how fast the desktop application responded to events started on the server. Again, iCloud Drive was among the fastest services together with Dropbox and Box.com. OneDrive and Google Drive were the two slowest services in this test.
The more I experimented with OneDrive and Google Drive and the more I thought that the desktop application for these two services is an afterthought. Something these companies had to develop but without putting too much love in it.
Conclusions
Despite the bad reputation that Apple’s cloud services have, iCloud Drive performed very well. It shows that it is tightly integrated with OS X as well.
Dropbox is the sync solution to choose if you don’t want to depend on Apple and if you need to work across platforms. Two of the biggest names in cloud computing Microsoft and Google also show that they decided to invest more on the server side of their solutions rather than on the OS X desktop application. I find OneDrive and Google Drive poor solutions even for basic use and I am not recommending them.
If security is your concern you should go with SpiderOak. The desktop application is not bad, but don’t expect the level of refinement that you can get with Dropbox. The speed is good too considering that every single file you upload is also encrypted before leaving your computer.
The pleasant surprise is Syncplicity. They give you a generous 10GB free plan and the service is backed by one of the big names in enterprise computing.
All in all the conclusion is simple. Stick with iCloud Drive if your life is in the Apple ecosystem. If you need more flexibility, need to share files, have access to past versions of your documents and not be tight with Apple, then go with Dropbox.
Share this:
Energy Impact for the most common sync services
In OS X 10.9 Mavericks, the Energy tab of is a nifty tool that I’ve played with a lot since the introduction of the new operating system.
In this post I’ll show you the peak Energy Impact and and Average Energy Impact for five of the most common sync services on the market: , , SkyDrive, Box and .
I have striven to set up the experiment so as to guarantee equal conditions for all services. I made sure that:
All four sync services contained the same files before starting the test
The services were idle
The same files/bundles/folders were uploaded to each service via Finder
New files were uploaded only after the previous ones finished syncing
To better simulate a real world scenario, between the tests I also performed some copy/pastes and file deletes, both from the web interface of each service and from the sync folder.
I tend to believe that this test has only academic value and somehow suspect that different users might get different results. It’s likely that the values I have collected are function of the processor my Mac uses, how many processes were running during the tests and some other parameter that I haven’t taken into account.
Nevertheless, you might find the results useful if you care about the energy consumption of the applications you use or suspect that something is not right in your Mac.
Results
The Average Energy Impact was simply retrieved from Activity Monitor within 8 hours of the end of the tests.
The Peak Energy Impact was retrieved by reading the maximum value of the Energy Impact column for each single sync service while uploading a specific file.
These are the results:
and as graph:
As you can see I have also plotted the Time To Start Syncing, which is – with enormous approximations – the measured time to see the sync service icon in the menubar to become active. To avoid adding useless digits, I approximated the time to the nearest second:
Somehow I thought that the Average Energy Impact was correlated to the Time To Start Syncing. That is, I though that the energy consumed was inversely proportional of the polling period (i.e. the interval at which each sync service checks the sync folder and remote server for changes). That is not really the case, at least in my tests. Obviously other factors affect the energy consumption.
What I found really peculiar was Dropbox’s behaviour. The peak Energy Impact was very high in all my tests, as if the service was doing something straight after adding a file to its sync folder. I don’t have proof of it but this might be related to the fact that Dropbox splits each file into chunks and calculates its hash before uploading it to its cloud service.
The case of OmniPresence
I’d like to dedicate a short paragraph to OmniPresence, the sync service created by OmniGroup for its applications (OmniOutliner, OmniFocus, OmniGraffle etc.). I use OmniGroup applications constantly, with frequent file changes and synchronizations from my Mac, iPad and iPhone. What I have never realized is how incredibly well optimized and low energy impact this sync service is.
On my Mac, the Average Energy Impact for OmniPresence is a low 0.10. Incredible. These guys know what they’re doing and the existing applications are there to prove it.
Conclusions
None of the sync services I have tested use Power Nap, which would dramatically reduce the energy impact on our Macs. I think that for a sync service Power Nap is not an option. That would likely impact negatively the sync times and responsiveness of the service.
As I suspected the most popular services also show lower energy impacts. This is good news for us because we all rely on them during our normal working day. Comparing the results of these sync services with other applications I normally use, I realized how low impact they actually are.
That is also reflected in the battery life of my MacBook Air. In non-scientific tests, I’ve never actually seen an appreciable decrease in battery life with one of the sync services running.
Share this:
Google Chrome and Drive for iOS updated
It’s a busy day for users trying to keep their iOS devices updated. Google has just pushed out updates for Google Chrome and Google Drive for iOS.
Google Chrome
Improved voice search:
Say what you want a d get results back without typing
Faster voice recognition with text streamed on the fly
Get answers spoken back to you with web results tailored to your questions
Faster page reloading
Pages reload faster even when the network is slow or unavailable
Stability and security improvements and bug fixes
Google Drive
Swipe through your images stored in Drive
View, create, reply and resolve comments in Google Docs files.
Share this:
Create your own private cloud with Transporter
by Connected Data is a solution to create a private cloud.
The target audience of Transporter is users who are concerned of the privacy implications of public clouds (Dropbox, Google Drive etc.) but that at the same time need to be able to access all their data from across the world.
Transporter reaches this goal by using a mini-server that backs up the data on your computer and makes it available through the Internet to any authorized device.
The great idea about Transporter is that you can link more than one storage device to extend the redundancy and capacity of your private cloud. By doing so you can:
Extend the capacity of your private cloud
Create effective off-site backups of your data.
For example, if you use two storage devices, one located in your home and another one at your sister house, then the data is intelligently mirrored across the two storage devices. If anything happens to the device in your house (theft, damage, etc.) your data remains still be available on the other device.
The company has also released two apps for iPhone and iPad so you can access your data while on the go.
Transporter starts at $ 199.00 and you can buy it .
Share this:
Google Drive for Android updated with scan functionalities and OCR
Google today has announced an important update for its Android Google Drive app:
The updated Drive for Android app also gives you to a way to keep track of important paper documents like receipts, letters, and billing statements. Simply click “Scan” from the Add New menu, snap a photo of your document, and Drive will turn the document into a PDF that’s stored for safekeeping. And because Drive can recognize text in scanned documents using Optical Character Recognition (OCR) technology, a simple search will retrieve the scanned document later.
Google is building something very solid with Google Drive on Post-PC devices. I am pretty sure we’re going to see this update on iOS soon.
, the leading virtualisation solutions company, has started beta testing its own version of drive in the cloud.
Contrary to similar solutions developed by Dropbox, Google and Microsoft, VMware Project Octopus will target the enterprise market by delegating to a central entity what file can be shared by whom and on which devices depending on rules defined at corporate level.
One of the main criticism of the solutions currently on the market is that whoever happens to have a certain file can easily share it as long as they have an Internet connection.
, have enhanced its cloud offering by releasing a native Mac and Windows software to access its product.
The battle is becoming even more interesting and we customers can only benefit from it.
[Edited on 22/05/2013 – In case you are interested in a speed test of different solutions, after you finish this post, you can go right here. There is a speed test of Dropbox vs Google Drive vs Box vs SkyDrive vs Google Cloud Drive]
As in the other post, let me print this disclaimer so that nobody can complain of the lack of scientific methodology behind my tests:
I acknowledge that this testing methodology is far from being scientific.
My setup and test methods might have affected the results and other people might get different values even opposite from what I am sharing in this post.
The results posted on this blog are for personal use only and do not constitute a definitive proof on what service is faster.
More tests, using different scenarios, different networks, and a systematic way to time events are needed to calculate average times, variance values and to draw some more informative conclusions.
Methodology and configuration
I have only used my MacBook Pro 15″ with both Dropbox and Google Drive clients installed on it.
The tests measure the time it takes to upload four different types of files to Dropbox and Google Drive. For the tests I used these files:
ZIP file, 23MB.
XML file, 9MB.
MP3 file, 10MB.
One folder containing folders/subfolders up to three levels and a total of 2663 files and a smaller one with 673 files. More on this later.
My test methodology simply involved carrying out a Copy Item and Paste Item operation between folders while measuring the time using my iPhone stopwatch.
The errors that you invariably introduce when you execute tests in this way can be so big that I wisely decided to not consider fractions of a second in my measurements. All data are correct to the nearest second.
To avoid that other software might influence the upload speeds, before running my tests I quit all browsers, mail programs, RSS readers etc.
Test results with single files
I have collected the results in this table:
and in these three graphs:
As you can see from the results, in this David and Goliath battle Dropbox performs quite well.
Dropbox splits the file you upload into chunks and calculate the hash of each of them. If a chunk with the same hash already exists on the cloud, Dropbox doesn’t upload it but simply match it with the existing one.
It looks to me that Dropbox have offset their limited resources with smarter algorithms whereas Google has simply given us a hard-drive in the cloud so to speak.
This will probably be the recurrent theme in the battle between the two services. Google going for sheer size, speed and integration with its other services whereas Dropbox will likely add small, elegant features to differentiate its product from the competition.
Test results with folders
I was curious to check the behaviour of a copy/paste operation of some useless temporary folder with a bunch of small files in them.
In my experience, different systems behave completely differently when it comes to deal with a large number of small files.
Somehow my curiosity has rewarded me, or maybe has just added a question that I don’t have the answer to. This is what I’ve found.
I first copied to Dropbox these two folders:
Folder A with 2663 temporary files in it and a structure with three level of subfolders.
Folder B with 673 temporary files in it stored directly under the main father folder.
Dropbox behaved quite normally, it took about 3 minutes to copy Folder A and about a minute for Folder B. Quite ordinary.
When I repeated the same test with Google Drive I was surprised to discover that Google Drive managed to complete the job only after 15–20 minutes (I stopped measuring the time because I though there was no reason to continue the experiment). To make sure that I wasn’t suffering from some broadband problems or an unavailability of Google Drive, I repeated the same test other three times at different times of the day getting exactly the same results. Why?
Why would Google Drive have such problems uploading this type of data?
If somebody has any idea I’d like to hear it and I would also be curious to know if you can reproduce the same behaviour. For your information I attempted to copy
As you can see from the tests that I carried out, both solutions return similar results. In case you upload the same file multiple times, Dropbox is dramatically faster thanks to the smart use of algorithms that make sure that file chunks with the same hash are not uploaded twice.
The obvious conclusion is that if you base your decision to use either Dropbox or Google Drive on pure speed only you will not get a clear answer. Dropbox in particular circumstances can be way faster than Google Drive despite not having the massive data centres Google have all over the world.
Google can entice us in many ways to use its Google Drive but speed is not one of them. Obviously the big advantage for Google is the integration of Google Drive with the other services available. It is definitely in that arena that the real battle for the best consumer cloud solution will be fought.
For instance, just two days ago Dropbox introduced the possibility to upload pictures directly from your camera and I am sure that we’ll see many other innovations coming our way.
Share this:
Does Google Drive signal a change in Google’s revenue strategy?
The real questions is what type of company would you trust with keeping your data in the cloud?
Microsoft is a software company, which sells it’s [sic] products and services to customers for money. You are the customer.
Apple is a hardware company, which sells it’s [sic] products and services to customers for money. You are the customer.
Dropbox is a web-service company , which sells it’s [sic] services to customers for money. You are the customer.
Google is an advertising company, which sells it’s [sic] services to customers for money. You are not the customer — advertisers are customers. You are Google’s product.
Now, I wonder if Google is trying – with Google Drive – to timidly diversify its revenue stream by making customers pay for the service they provide instead of being the latest way to mine your data so that they can targets their ads more precisely.
I admit that this question is provocative but maybe there’s something behind my reasoning:
This is not a beta product like so many other Google products in the past. With the exception of the iOS client (short sighted anyway) this is a finished product, not a half baked one.
The time between the announcement and the roll out has been very short. Again, this is not like Google TV or Google Music. Google wants people to use it, now.
Google Drive has been enabled for Google Apps for Domain at the same time as normal Google Accounts. This is unusual for Google, Google Apps customers are always behind when it comes to get the latest Google services. This signals that Google is targeting the service to companies as well in a moment when Google Apps for Business is doing pretty well.
Most important . For a basic 20GB drive now you spend almost six times more than what you used to spend with the old pricing structure. Google had the capability to maintain the old prices and massively data-mine your files but instead chose to increase the prices quite a lot. Why?
Google realises that having 96% of your revenue from ads put them in a vulnerable position especially in an worldwide weak economy hence the need to find other sources.
Anyway, this has been a fun post to write, maybe far fetched but fun. What is your take on Google Drive and how does it fit in Google’s business plan?