Cloud Storage: We need more?
I would like to write this report on the world situation of Cloud Storage services, in light of Apple’s beta release with iCloud.
First of all, let’s say what Cloud Storage is.
It is an online service, through the internet, of a data storage system, which translates into our personal computers or mobile devices, in a kind of additional drive where we can save our data “unlimitedly”, knowing that if we miss the PC we will certainly not lose those data. It is also often used to synchronize folders between home and office, for example, or synchronize the address book, calendar, etc etc. It allows data to be shared with other users on the network, for example it is very useful as a collaboration tool between multiple locations. It is very useful for relieving application servers, taking charge of the publication of content.
This service to be defined as Cloud should be based on the logic of pay per use, be self-service, measurable, elastic, and be highly reliable and perhaps distributed in multiple sites around the globe.
Let’s make a small list of what’s currently out there:
- AWS S3 – An infinite size http/https protocol storage, distributed and redundant countless times, is used in authentication through an account on Amazon Web Services, not very fast, but with very high reliability. It is more than 5 years old, it is the father of many other storage services including….
- DropBox – Which remembers about three years and uses S3
- Amazon Cloud Drive – recent, I guess it relies on S3, think that many columnists on the net at the release of this service declared that Amazon was also jumping into the Cloud
- Microsoft Azure Storage – One of the 4 cores the blob is conceptually similar to S3, perhaps more performing, but it is also younger, it will be a couple of years old
- SkyDrive – Microsoft Live Platform Storage Drive with 25GB Free
- Ubuntu One – is an online storage service offered by Canonical for its Linux Ubuntu, also available for Android, iPhone and soon also for Windows, as well as the Amazon Cloud Drive, are advertised to be used with music files, photos, videos, in short, for entertainment
- Google App Engine Datastore – I would compare it to S3 and Azure Storage Blob, they are all three programmable by scripting that interface with the protected webServices of their endpoints. Just think that these three services, as it happens, can generate an infinite number of Cloud Storage offers where the limit is only the imagination of entrepreneurs and programmers. For example, classes in many languages are public to do basic operations, such as upload, delete, get. For S3 the classes are also public for Android and iPhone, you just have to search on search engines and you will find an infinite number of them. I remember that already in 2007 I had a public C# code to create an additional drive for windows always connected with S3, this also had a buffer for latencies and disconnections
- RackSpace CloudFiles – Compares to S3 and how it can publish content via CDN
- GoGrid Cloud Storage – A storage that is generally used to manage the data of the various servers of their Cloud Computing platform, but also accepts classic network protocols such as Samba/CIFS, SCP, FPT and RSYNC
- Nirvanix Cloud NAS – Nirvanix’s storage is one of the oldest, it has been on the market for many years, they have a very solid experience and recently their offerings have diversified and increased.
There would be many others to list, but I would now like to say that there are also many open source projects to be able to create your own Cloud Storage with classic servers at “home”, projects that are often the matrix of these great ones that we have listed. For example:
- Apache Cassandra, for structured data, used by facebook
- OpenStack object storage,
- Ceph,
- GLuster,
- pNFS,
- Dynamo, the S3 project
- Voldemort, a project used by Linkedin, always key-value
- Lustre, distributed filesystem
- Google File System
- Global File System di RedHat
- Hadoop distribuited file system