One of the strengths of Komprise is what we call Transparent Move Technology™ (TMT). We recently published a white paper overview of TMT and I wanted to dig into some of the common questions we hear when we introduce TMT. Who better to go to than Kumar Goswami, Komprise CEO and Co-founder. In this two-part interview, we’ll talk about the unique (and patented) approach of TMT and the benefits of not being in the hot data path.
What gets you most excited about TMT?
KG: We focused on making it simple for users to embrace the potential of data management because Komprise requires no change to users, application workflows or processes. With our patented Transparent Move Technology, users and applications can continue to access moved data exactly as before. Komprise provides two options.
- First, we can move data transparently so that users can still list, search and access moved data from the same location as they did before the files were moved. TMT allows IT to specify policies for tiering and archiving data to ensure cold data is continuously and systematically moved from the source without disruption to users, applications or IT data protection workflows. Komprise puts IT in control to operationalize the solution for maximum savings. TMT does this without any stubs or agents, using just open standards.
- Second, Komprise also provides a policy-driven way to “lift and shift” data completely off the source to file and object store targets. Nothing about the file is kept on the source. This approach does not provide user and application transparency, but it may be ideal for data-heavy projects where users will likely no longer access the data. For such cases, the moved data can be accessed directly from the target or by mounting Komprise as an NFS or SMB share. In the latter case, users can list and access the moved files or drag and drop these files into other sources easily and without requiring IT intermediation. Note that this is possible even if the files are moved to an object store or the cloud. You get both file and object access regardless of the target.
What are some of the common issues you see with proprietary tiering solutions provided by storage vendors?
KG: Proprietary data tiering solutions are implemented at the block level and not at the file level. Here are the issues that result:
- They access and interact with cloud storage using the same complex block management mechanism and protocol designed for their internal storage platforms. While fetching and writing to small blocks within your own storage might work, it does not work with the cloud because the cloud has much higher latency and you must pay to read and write objects leading to unnecessarily high costs and poor performance. As a result, this block-based approach is highly inefficient, leads to greater cloud egress costs, and potentially slows down overall performance of the storage unit.
- With data stored as proprietary blocks you also can’t access the data from the cloud. You cannot take advantage of cloud-native AI/ML workflows nor advanced security features such as S3 Object Lock to protect your data from ransomware. Essentially, block-based storage pool tiering creates a digital landfill of proprietary blocks in the cloud.
- Proprietary solutions don’t reduce the backup and DR footprint, especially for third-party backup tools, leaving much of the savings from tiering cold data on the table.
- Finally, these solutions don’t provide any analytics so you cannot see all of your data, regardless of where it’s stored. This means you can’t holistically plan your data management strategy and approach.
Komprise TMT, unlike storage array tiering and cloud storage gateway solutions, works at the file level and doesn’t have these issues. It works equally well with cloud storage as well as on-premises object stores.
Does TMT work with backup solutions?
KG: Yes. Komprise TMT works with all standard backup solutions and does not rehydrate the data, thereby reducing the backup footprint and ensuring significant cost savings.
TMT simplifies the movement of cold data to less expensive storage. You can literally make your data moves with a click of a button, without any disruption to users, applications, data protection workflows, or access to mission-critical hot data.
While some storage vendors provide proprietary tiering solutions for cold data, these don’t provide analytics or a granular policy that can be customized for each share. And if you use them with a third-party backup solution, all archived data may be rehydrated. You’ll need to maintain the full capacity of your existing data to rehydrate all archived data, thereby eliminating any cost savings!
Why is not being in the hot data path such an advantage for TMT?
KG: Many data management solutions sit in front of your primary storage and divert requests for the cold data to the tiered location. This affects the performance of hot data since it introduces a middleman. A “traffic cop” now directs data access, which is a tremendous risk for all your company’s data. A failure in this system creates an access nightmare and when access to all of your data is lost, you’ll hear far worse than honking horns.
This approach also requires that scaling be based on hot rather than cold data access rates. Since hot data is 99.999% of your data access, this means that the so-called “man-in-the-middle” device must be able to handle the massive hot data traffic requests. As this traffic increases with your data growth, the “man in the middle” must scale accordingly and still handle access spikes. If you don’t plan accordingly, you’ll decrease performance on your new flash storage. These types of solutions—and their issues—have been around for years and have resulted in customer defections due to high cost, decreased performance, and risk.
In part two of my interview with Kumar Goswami, we’ll talk about how Komprise TMT is different than stubs and the benefits of direct data access.