Cloud data access: Transport retrieval is not easy

Before putting data into public cloud storage, IT departments need to consider how to take the data out.

Simple data access is straightforward, Taneja Group analyst and senior consultant Jeff Byrne said. Most cloud storage providers support a web-based schema-based Table Property State Transfer (REST) application Interface (API). Some also support traditional block and file based data, and cloud storage gateway providers will assist customers in accessing data in primary storage cloud.

But Byrne says customers often rely on their own power if they want to transfer data from one cloud provider to another or get their data back into the enterprise. In this podcast, TechTarget senior journalist Carol Sliwa a dialogue with Byrne, and Byrne talked about data access mechanisms for major cloud storage providers as well as some of the issues associated with transmission.

TT: How much work does it involve if a business data is already stored on a cloud provider and need access? In general, how do you specialize the process based on time and ease of use?

Jeff Byrne: The current approach is straightforward, and the ease of access and time required depends on the type and format of the data, that is, the level of service that the user has for the specific storage type. For example, developers who want to write new applications for cloud implementation will typically use Cloud-based object storage services, such as Amazon S3. To access its data, in S3, developers use the Web Services API in principle, and now most of the cloud storage APIs, including Amazon S3 and OpenStack Cloud, are compatible with rest. Therefore, this is a generic API.

To continue our Amazon S3 case, the developer accesses and operates its object data in the S3 bucket, and can use the "get" command to access specific data objects stored in S3. Basically, a developer who is familiar with the rest-style API can quickly access its data, which is very easy if the data belongs to hard disk storage. For example, in a developer's notebook or in a data center.

There are many quick access cases. But if you talk about data in other areas, such as archiving data, the process will be longer. For instance, a new Amazon service, probably launched in the middle of last year, is called "Glacier". As the name suggests, accessing data on glacier is time-consuming because it is designed to archive storage services. The assumption is that user data is not accessed frequently. Therefore, in this example, it will take 3-5 hours to get the data.

TT: How different are the data access methods between the primary cloud storage providers?

Jeff Byrne: Data access methods vary from provider to vendor, based on the APIs they support and the data types that users store. Again, if you're talking about developers using cloud object storage cases like S3. APIs are available on most cloud storage sites. But there are other protocol support. For example, Azure still supports soap (Simple Object Access Protocol). There are many Web sites that support WebDAV.

And not only are these newer web-based protocols available, but many providers support more traditional chunks or file based data. Nirvanix Cloud Storage Receptacle and IBM SmartCloud Enterprise, for example, support file-based APIs, such as NFS and CIFS. There are also vendors that support block based access methods, such as iSCSI. IBM and Cleversafe are examples of this.

In addition, I would like to mention the cloud storage gateway vendors. Ctera and Storsimple are two examples. Storsimple is actually part of Microsoft now. What these vendors do is provide local applications that facilitate access to file-based or block-based data in the primary storage cloud. I really think that these vendors are filling most of the important needs of businesses that want to take advantage of the cloud, low cost, universal access, and so on, but they can't afford to redesign their apps and create new cloud formats for their data. These gateways allow them to do so, basically by moving existing data into a device, then being translated into the cloud in a rest-based format, and then accessing the data from the local device.

TT: How hard is it if the business needs to change and transfer data from one cloud provider to another, or get the data back into the enterprise?

Jeff Byrne: When users want to put data into a cloud site, there is a whole set of services there. Almost every cloud storage provider provides a degree of raw data seeding services or transportation services or, in any case, access to data. But when it comes to taking the data out, it's surprising, but most users do it themselves. So users need to focus on a lot of things. The biggest problem is that the cloud industry still lacks interoperability standards, which means that businesses that need to take data out must ensure that their new providers support the data formats they are using, assuming they are moving to new providers.

There are tools available, traditional file transfer tools such as FTP, users can be used to transfer data, but not good to connect to the cloud data. I think if you're tired of using these tools, you'll find that the process is not just time-consuming, it's cumbersome, because you have to write scripts for these types of tools, and then if you do it yourself, the data transfer itself becomes delayed and lacks security.

Therefore, if you are transferring very large amounts of data and cannot afford to lack access and compromise on these data, you may be very much in need of a third-party migration solution. There are a number of solutions available. Some are open source, some are commercial, but not too expensive. For example, a very prominent third party, called Cloudberry Lab, its tools called Cloud Migrator, support Amazon, Azure, Rackspace and so on. Products like this allow you to make direct, basic data transfers between buckets, such as in cloud sites like S3 or from one cloud provider to another. Also, these service tools are safe and automated, for example, they do not require software to be installed on the local server. Therefore, quite without hindrance. Before you transfer data to a second cloud provider, you do not have to first shift it to your data center.

Other alternatives to mention, once again, are cloud storage gateway providers, because most gateways support major cloud storage providers, and they have the right tools to migrate data from one provider to another.

(Responsible editor: The good of the Legacy)