Cloud Computing, Chapter 7

Chapter 7: Cloud Storage

Storing data with a cloud service provider...

Overview:

If stored in the cloud, your data can be accessed from any location. There are hundreds of different cloud storage providers, ranging from general to very specific niches. Like flicker with pictures. All that is need for a storage system is a data server connected to the interwebs, most systems are made up of dozens or hundreds of servers. Due to maintenance it is necessary to store the data on multiple servers, without this data access couldn’t be guaranteed at all times. Cloud storage is used, often, as a safety net so that if something happens to their computer- like a hijacking-  the data is still accessible. Storage as a Service (SaaS) is when a third party rents space on their dataservers to end users.

Specialized Cloud providers are google docs (documents, spreadsheets, powerpoints), web email from gmail, yahoomail and hotmail, flickr and picasa specialize in pictures, youtube hosts user videos, and facebook and myspace are networking sites that allow users to post pictures and videos. Many of these are free, but some storage charges per GB of data. Securityt takes the form of encryption using complex algorithms to encode information, authentication through a user name and password, and authorization practices  where the client creates a list of people authorized to access the data. However the threat of hackers is always hanging in the air.

The reliability of the system, if you are unable to access the data stored due to a system crash, but with how many different services available those unreliable providers will have few clients and often aren’t around long. Cloud providers move and balance the information on their dataservers making the information close and easily accessed. Cloud storage is a great safety net for data in case of a disaster, the files are okay and backed up off site.

It is best, at this point in time, to used a mixed approach as cloud storage is still immature. It is recommended that the cloud is used for noncritical purposes. The most critical factors to take into consideration when thinking of moving to the cloud is price and reliability, try different services to find the one best for your needs. Outages are a possibility, so take that into account when deciding on a provider. Security and theft are risks with cloud storage, when the data is away from your own datacenter; make sure to encrypt your data and secure the data transit.

Providers:

(non comprehensive list)

Amazon Simple Storage Service (S3)-

Launched in 2006, S3 was designed to make wed-scale computing easier and can be used to “store and retrieve any amount of data, at any time, from anywhere on web”(142). It allows any client to use the infrastructure Amazon uses. It allow for an unlimited amount of objects to be stored, written, read and deleted. Each object is stored and retrieved with a unique key, they can be public, private, or for specific users. S3 was built to be scalable, reliable, fast, inexpensive, and simple. In order to do this Amazon made the system decentralized to remove single points of failure, autonomic so that individual components can make decisions using local data, localized responsibility, little or now concurrency control required, failure of components normal, parallelism, build small components that can then been built on for other services, simplicity.

S3 works using a “bucket system” where each object (up to 5GB) is stored with 2KB of Metadata, these objects are organized by “buckets” and are owned by an AWS account and identified by a unique key. Developers only pay for what they use!

Nirvanix-

Software file system stored on intel storage servers, they are expecting to add more server locations. The Nirvanix Storage Delivery Network (SDN) turns a server into network attached storage.

Google Bigtable Datastore-

Made public in 2008, Bigtable was developed to be highspeed, flexible, and high scalability. It is available as part of the google app engine. Developers have been cautious about using bigtable as they get locked into google. Cost wise it runs about the same as Amazon.

MobleMe-

Apple’s solution with email, contacts and calendar. They also provide ad-free web applications for a desktop like experience. Push keep contacts, email and calendars up to date no matter what you use to access the MobleMe.  20GB is $99 a year for an individual or $149 for a family, additional space can be purchased.

Live Mesh-

Microsoft’s solution, which allows PC’s and other devices to be aware of each other. Live mesh is made of a platform defining and modeling a user’s digital relationship among devices, data, applications, and people; hosted in Microsoft datacenters, software that enables local applications to run offline. The software, Mesh Operating Environment is available for XP, Vista, Mobile and OS X.