One of the first question that you should ask when moving to Google Cloud is how to protect your data. Given the increasing importance of information, data protection is more critical than ever. In this blog post, we will look at the pros and cons of four ways to protect data in Google Cloud.
Option 1: Persistent Disk Snapshots
- Native tool – Google Cloud provides native snapshot functionality and so this technology is easily accessible.
- Incremental forever – Google snapshots are incremental forever and so only changed blocks are stored. This can enhance the efficiency of data protection.
- Not application consistent – A key requirement for many customers is the ability to quiesce applications when creating snapshots to enable rapid recovery. Google snapshots do not typically include this functionality and so manual scripting or other similar activities may be required which adds to complexity and reduces backup reliability.
- TCO – Data retention requirements vary by company. Snapshots are typically best for shorter retention requirements; however, if you need backup of retention of weeks, month or even years, the costs of snapshots can increase rapidly thus delivering a very high TCO.
Option 2: Traditional Backup Applications
- Single platform for backup & recovery – While traditional backup products are optimized for on-premises instances, you can often use them in the cloud as well. This provides the benefit of a single application with a common interface.
- Flexible policy engine – Traditional backup applications provide robust and mature policy engines for data capture, retention and deletion. Extending this to the cloud enables a similar level of functionality in the cloud and on-premises.
- High RTO – Traditional applications perform incremental and full backups and so recoveries require a restoration of a full backup and related incrementals. Data can only be accessed after all the data has been copied and as a result recoveries can be a lengthy and I/O intensive process.
- High TCO – On-premises backup applications are typically optimized to run locally versus in the cloud. While they can run in the cloud, the costs of computing, memory and storage can often be higher than expected resulting in much higher costs to operate versus cloud-centric alternatives.
Video: Googles Speak About the Future of Data
Option 3: Hyper-converged Appliances
- Scale – The concept of hyper-converged appliances incorporates the idea of a physical or virtual appliance that includes all the required storage, compute and memory. These solutions typically scale efficiently and provide simplified ease of use on-premises and in the cloud.
- High TCO – These appliances typically rely on deduplication which requires significant compute and memory. To make matters worse, you usually need at least four virtual appliances to provide the appropriate data protection and so the cloud costs can get very expensive.
- High RTO – The recovery process for hyper-converged appliances is similar to that of traditional backup applications and so the process can be lengthy.
- Limited application support – Many customers look to new open source databases like MongoDB, MySQLor PostgreSQL in the cloud. Hyper-converged offerings typically have a reduce compatible matrix as compared to traditional applications. Thus they may not provide the performance and efficiency for these databases.
Option 4: Next Generation Approaches
- Application integration – Newer protection solutions will typically support the common open source databases which simplifies implementation in the cloud.
- Data efficiency – While traditional protection offerings rely on periodic fulls, newer technologies leverage more advanced strategies like incremental forever or deduplication to reduce data movement and storage.
- Cloud optimized – Next generation approaches natively embrace the cloud and leverage inherent cloud features like object storage. This is different from traditional solutions that are more on-premises-centric.
- Newer vendors – Typically next generation solutions come from newer vendors who are bringing new strategies to data protection.
- May only work in the cloud – Users often have a heterogeneous environment with some pieces running in the cloud and others on-premises. Some solutions only support the cloud which can add complexity as it could require separate solutions for the cloud and on-premises.
Whitepaper: Learn how Actifio helps protect data in Google Cloud