A Comprehensive Look at Data Archive Strategies

PACS Training - IT Basics 101
Let’s Explore the Different Data Archive Strategies Available Today

We are living in a digitally advanced era where every organization is generating, processing, and storing tons of digital data. Organizations that deal with large flows of digital data usually set up strong data management policies to ensure they manage the data rightfully. Data management involves different processes and a well-planned approach to handling data. One such process is data archiving where the less frequently used data is stored in a low-cost storage medium. In this course module, we will talk about data archive and the different strategies around it.  These strategies affect the performance of image load times on PACS and VNAs.

What is Data Archiving?

Data archiving is a process of transferring less frequently used data to low-cost storage repositories. It is meant to reduce costs by avoiding storage of such data on main storage, while still retaining the data for future analysis or regulatory compliance. Archived data is stored in cold storage ties to enable long-term retention and cost-effective data storage. 

Data Archive Strategies

Data archiving is not just limited to storing less accessed data in low-cost storage mediums, it also involves well-planned strategies on what data to store, how to transfer data, how to make the data accessible, etc. Through thoroughly prepared data archive strategies, an organization can ensure long-term data retention in the most efficient way possible. Let’s explore some of the key terms under archive strategies:

Online

Online storage refers to immediately available storage that is quickly accessible.  Online storage is typically non-removable storage media that is instantly available.  Examples are solid state drives, spinning disk drives or redundant array of independent disks (RAID).  In the PACS world, it is useful to have recently performed radiology studies and the priors available in online storage.

Nearline

Unlike online storage, nearline storage is not immediately available.  Nearline storage is the storage outside the computer (mostly removable media) but provides quick access to the stored data when needed. Nearline is the phrase built from two words, i.e., near and online. Basically, it refers to any storage infrastructure that resides between offline and online storage sources. 

Organizations are using nearline storage devices to store archive data and access it on-demand without involving any human intervention. For example robotic arms.  Common nearline storage products include magnetic tape, magnetic disk, compact disk, optical disk, etc. Nearline storage devices are not attached to the computer, so they are excellent for storing archive data and protecting it from security and data loss threats. 

Cloud

With the growing concept of cloud storage, storing archives online is considered the easiest way to store and access data from anytime, anywhere. It also streamlines data management and frees the organization from maintaining on-premises data management resources. 

The leading cloud services are also offering dedicated online archive offerings. For example, Amazon S3 Glacier offers different storage classes meant for data archiving in a low cost and secure way. Its basic package starts at $0.023/GB for the first 50TB/ month. In short, archiving data online is one potential choice for organizations today.

Data Throughput

Data throughput reflects the amount of data that is delivered successfully from one position to another in a specific period of time. It takes into account the network speed, latency, packet loss, and other factors then presents the actual amount of data that reaches the destination. 

It is usually measured in gigabits per second (Gbps), megabits per second (Mbps), and bits per second (bps). For example, let’s assume that a dataset of size 100 bytes takes 1 second to transfer from one system to another. So, the data throughput between the two systems will be 800bps. Since 1 byte is equivalent to 8 bits, that’s why it’s 800bps.

Methods for Storage Management

Storage management refers to the processes and software that are meant to enhance the performance of data storage mediums in compliance with the organization’s policies and government regulations. Some of the common methods of storage management include: 

  • Virtualization: Storage virtualization is pooling physical storage from two or more storage devices into a central virtual storage device. All the storage devices that are part of the pool remain unseen, thereby making the virtual storage one comprehensive physical storage. The input/output requests are handled by virtual storage software that seizes the requests from physical/virtual machines and then sends the requests to the targeted physical storage location. 
  • Replication: Storage replication is the process of making one or more copies of the data and storing them in different locations. Organizations actively carry out storage replication to ensure quick recovery of data in case of any data loss incident.
  • Mirroring: Storage or data mirroring is the process of making an exact copy of the data and storing it in a different location whether on-site or off-site. It is practiced mostly in cases when there is a need to have exact copies of data at multiple locations. 

Other than the above 3 methods of storage management, an organization can practice other methods too, such as traffic analysis, network virtualization, process automation, memory management, etc.

Storage Metrics

Storage metrics are useful statistics about the data storage management system that can reflect general performance, detect unusual storage consumption, pinpoint hidden issues, etc. Some of the common storage metrics are as follows:

  • Capacity – Reflects the storage space, including the total, used, and free space. 
  • IOPS (Input/Output Operations Per Second) – Reflects the number of read/write operations the system can handle.
  • Latency – Reflects the time it takes for the storage to complete the transfer.
  • Throughput – Reflects the amount of data that is delivered successfully from one position to another in a specific period of time.
  • Queue Length – Reflects the number of input/output requests pending at a specific time.

There are plenty of other storage metrics that organizations monitor to ensure that the storage management system works at the optimal level.

Final Words – Data Backup vs. Archive

Although the working principle of data backup and archive seems similar, they have many main differences. Data backup is the copy of the organization’s current active operational data. It involves data that is accessed and used frequently. Moreover, a data backup is a replication of the data, which does not affect the original files. Besides that, data backup is recoverable instantly.

On the other hand, data archive is the data repositories of those datasets that are not important but should be retained for the long term. Archived data does not change often and is not accessed frequently. They are the original data files, not the replication. Moreover, they are stored for a much longer duration compared to backed-up data.


How to Become a PACS Analyst

Step by Step Guide on what to learn and how to enter the field of Imaging Informatics