Limitations of Current Approaches
A snapshot is the state of a system at a particular point in time. It is a well-known industry term. We expect some sort of snapshot functionality from every modern computer system. Snapshots provide us with basic protection from data loss due to accidental or malicious deletion of data. They also improve application availability, accelerate recovery, simplify backup of large volumes of data, and reduce the overall exposure to data loss. Rolling back to a previous version of files, directories, or entire volumes is vastly simplified. It can also significantly reduce backup windows. Snapshots has become so commonplace that we rarely think about it beyond its basic functionality.
But not all snapshots are created alike. There are a number of different vendor implementations as well as open source alternatives that all do things slightly differently. In the previous paragraph we listed the goodness that snapshot technologies bring. Let’s examine some of the shortcomings or overlooked areas next.
The challenges of managing unstructured data is a complex problem; and that complexity seems to grow faster and faster. Can existing snapshot technologies keep up with such unrelenting growth? There are several questions we should ask with regards to snapshots. Are snapshots subject to the limitations imposed by infrastructure growth, such as sprawl and silos? How do you control costs and effectively manage larger and larger amounts of snapshots? What if those problems could be solved in a way that not only improves manageability but also reduces infrastructure costs?
Data sprawl is a major storage ill. The demand for more and faster systems to hold data creates sprawl. We need more systems to provide more storage, faster and different systems to adapt to unrelenting data growth coupled with ever-tighter timelines on smaller budgets. This leads to a data management nightmare, where snapshots from multiple and incompatible implementations have to be separately managed. Snapshots, thus, inherit the ill of data storage sprawl.
This dilemma is compounded by the fact that different snapshot implementations require different configuration options. Frequency and retention schedules have to configured and managed separately. Each implementation has different limitations and idiosyncrasies that requires learning and maintenance of skills. It adds up to more and more touch points which, of course, consumes more and more of your time. Additionally, silos disallow the ability to leverage snapshots for reporting and analytics in a centralized fashion across systems. There is no economy of scale in an endlessly sprawling siloed world of unstructured data storage in a heterogeneous infrastructure environment.
Snapshots Consume Storage
Long term snapshot retention can consume large quantities of expensive primary storage. Storing all your cascading (hourly, daily, monthly, and yearly) snapshots on expensive primary storage is great from an availability standpoint. But do you really want to pay the price of retaining your long-term monthly and yearly snapshots on costly high-performance storage, such as flash? Eventually, you may have to acquire more primary storage capacity to make room for your long-term snapshots. That is an expensive proposition that provides no questionable value to IT or end-users. Your storage vendor will love you for it; but your pocketbook will not. Wouldn’t it be better if you could somehow offload your long-term snapshots to a lower cost tier?
The Hammerspace Approach
What if you could manage those snapshots in a way that eliminated or significantly reduced sprawl, silos, and storage consumption? Hammerspace is a next-generation data management solution that frees data from its underlying infrastructure. Managing data effectively puts an end to sprawl, silos, and out-of-control costs. And snapshots are an important part of that picture.
Curing Data Management Ills
Hammerspace eliminates sprawl and silos; and gives you the ability to choose the optimum performance vs cost strategy for your data. Hammerspace believes that the best way to achieve those goals is to centrally manage snapshots across your infrastructure, and throughout their life-cycle. Hammerspace has the ability to integrate with different vendor storage APIs. By doing so we can leverage native file clones across heterogeneous storage technologies to generate our own snapshots. This, in turn, allows Hammerspace to centrally manage snapshots from share-level down to single files, with granular controls, throughout their life-cycle.
Snapshots can be moved or copied to another storage volume. The natural target for this is a NAS or Object Store. And you are not restricted to one or the other. In fact, you can make as many copies as you wish to as many volumes as you desire. Snapshots are deduplicated and compressed when moved or copied to an object store, minimizing your footprint.
Hammerspace snapshots that have been backed up to another volume provide automated disaster recovery for Windows, Mac, Linux, VDI, and Hypervisor clients. If a primary volume becomes unavailable, Hammerspace will automatically recover directories and files if a snapshot has been previously copied to another available volume. Hammerspace snapshots, thus, improve availability as well as durability.
With Hammerspace you manage the placement snapshots on storage infrastructure that best meets your goals for availability, durability, and cost. Traditional storage solutions leave you with few options to cost-effectively manage availability and durability. Snapshots end up consuming valuable and expensive primary storage throughout their entire life-cycle. If you want to improve durability, you have to replicate to an identical storage array. The more you replicate, the more costly primary storage you end up consuming.
Hammerspace makes snapshots usable across multiple geographically dispersed locations, without having to replicate a full copy of the data to the other locations. Hammerspace supports up to 2048 snapshots per SMB share or NFS export. This combined with the ability to offload older snapshots to a more affordable tiers of storage provides snapshots for many years to come at optimal cost savings.