Continuous Data Protection (CDP) – Almost Perfect Data Backup

The degree to which businesses depend on data is increasing almost every month. This brings inevitable consequences, including the need to invest in advanced backup solutions.

Data has become a priceless commodity in recent years, something which, unfortunately, has become evident during the COVID-19 pandemic. Data is often the basis for maintaining customer relations, improving patient services, or providing a whole host of public services. However, the dependence of business on data also has negative implications. Almost any interruption or loss of data leads to severe financial losses and company bankruptcy in the worst-case scenario. According to Veeam’s Data Protection Trends report, as many as 95 percent of companies have experienced unplanned downtime in their lifetime. The study found that an hour of downtime for a high-priority application costs $67.6k. Meanwhile, for a case with normal priority, the cost is $61.6k. Add to this the deep cracks in an organization’s image following a cyber-attack or outage, which are difficult to quantify.

Backup System (High) Expectations

Managers are aware of the risks around them and are looking for solutions to prevent a possible disaster. Depending on the company profile, their expectations can be very high. Sometimes, business departments want IT systems to run 24 hours a day, and in the event of any failure, the infrastructure is to be back up and running in the blink of an eye. However, bear in mind that recovery speed comes at a price, and not every business needs express recovery. However, the fact is that an organization’s return to regular operation depends on the last backup created. Inadequate data protection can set a company back by hours and sometimes even weeks, thus reducing employee productivity. This is why the right, informed choice of backup systems becomes so essential. You should choose products that are tailored to the current needs of your company or institution.

What is CDP?

Advocates of instant action increasingly refer to Continuous Data Protection (CDP). Despite appearances, this is not a recent invention but a more than two decades old technology. It was patented in 1989 by entrepreneur Pete Malcolm, who defined CDP as a backup system in which a copy of every change made to a storage medium is saved as it occurs. What has changed since then, and can a discovery made over twenty years ago be a cure for the most significant data protection pains?

RPO Very Close to Zero

Entrepreneurs usually look at price when choosing a backup solution. But they also take into account other criteria such as the frequency of backups as described by the Recovery Point Objective (RPO), the maximum acceptable time needed to recover from a disaster as defined by the Recovery Time Objective (RTO), the extent and type of support offered by the manufacturer, and the ability to monitor backups. The deciding factors for CDP are slightly different. This high-end solution is designed to protect applications for which no data loss is allowed, so price plays a tertiary role. More on RTO and RPO you will find in our recent blog-post Data Backup and Disaster Recovery for Small and Medium-Sized Businesses (SMEs).

CDP differs from traditional methods that back up files at a specific point in time. There are no backup schedules with this technology, and the data is stored in a different location, usually on a different server. This means that organizations can restore data theoretically from any point in time. In ideal continuous data protection, the RPO is zero.

There will be virtually no loss for high-performance environments, while for less efficient environments, it may be several seconds. In practice, the RPO for banks is counted in seconds or even milliseconds. In contrast, the requirements for RTO designation are less stringent. In this case, the target recovery time should be defined in such a way as to avoid the unpleasant consequences of interruptions in the continuity of IT infrastructure or applications.

The Recovery Time Objective depends on the industry, the specifics, and the company’s size. A small online shop selling niche products can afford an RTO of tens of hours. But for a large e-commerce company, a significant financial institution, or a manufacturer, losing business continuity for 24 hours can be catastrophic.

CDP technology involves capturing continuous changes in data in real-time using various techniques. CDP-based solutions can offer detailed information about the recovered objects, ranging from failure-consistent images to logical objects such as files, messages, logs, and database files. CDP needs less space for image files than traditional backups in some situations. Most continuous data protection solutions record differences at the byte or block-level rather than at the file level. For example, if 10 bytes of a 1 GB file are modified, the system will back up only the modified byte or block.

In contrast, traditional incremental and differential backups typically back up the entire file. An alternative to CDP may be snapshots, allowing similar effects, but this impacts system performance.

Snapshot, Mirroring, or CDP

Professionals see CDP as the most advanced and comprehensive option for immediate data recovery. However, alternative technologies similar to CDP can provide a high level of security with slightly less complexity and therefore lower cost. One example is mirror backup, which, like CDP, stores every change to the source data. However, the key difference is that continuous backup allows you to go back and recover older versions of files that have been changed. In the case of mirroring, there is no such option. Another top-rated solution is the snapshot, an instant virtual copy of a given volume. A snapshot freezes an image of stored data at a specific point in time, which can apply to individual folders and entire file systems or databases. This is the state of the storage system captured at a given time. Preserving the system’s state makes it possible not only to recover data in the event of a disaster but also to restore it to known operating points. This is helpful in various scenarios. For example, snapshots can be used to go back in time before installing and uninstalling software, updating a system, or upgrading hardware components. They are also helpful for developers when testing and validating application code on virtual machines. However, snapshots are not a completely reliable solution as they do not protect data in the event of a disk failure. In addition, creating an excessive number of snapshots can lead to reduced system performance.


What speaks for CDP? Most solutions offering CDP provide an orchestration component. Such a component allows pre-planning for the recovery of critical groups of applications. In addition, you can also set SLAs at the application level and specify which systems have priority. It is also worth noting that CDP “doesn’t touch” the VM from the hypervisor’s point of view, which is another big advantage. The CDP protocol does not use a hypervisor or “snapshots” of storage.

In addition to the benefits mentioned above, CDP has several drawbacks. Firstly, it is not a cheap solution, and in many cases, there is no economic justification for using it. In certain situations, this form of backup can limit network bandwidth, especially when it involves multimedia and CAD (Computer-Aided Design) environments.

Paweł Mączka Photo

text written by:

Pawel Maczka, CTO at Storware