Organizations of all sizes are dealing with explosive growth in unstructured data. Whether the organization is large or small, they need to store and manage their unstructured data sets cost-effectively. They need an object storage system and software that turns data into an asset without breaking the budget.
The problem is that most solutions focus only on the enterprise. Many object storage vendors won’t even talk to a potential customer unless they have more than a 100TBs of unstructured data to store. Even large environments that DO have over 100TBs of unstructured data are not likely to move that 100TBs to a new vendor’s system at the point of initial implementation. The object storage market needs a “start small” option so that mid-sized data centers can reap the benefits of object storage and so enterprises can grow into the solution.
The Challenges with a “Start Small” Object Store
IT professionals face two challenges when attempting to “start small” with object storage. First, most, if not all, object storage systems are scale-out in nature. A scale-out storage system typically requires three or four nodes for the initial implementation. The four node requirement means that IT needs to allocate enough rack space to hold four servers, integrate those servers into the network and create the initial cluster. Second, most object storage solution vendors sell their solutions as two distinct parts. The customer purchases the software from the object storage vendor and then purchases three or four servers required to host the software from another vendor, or tries to repurpose existing servers. Then the IT professional still needs to install and configure the software to work with the hardware and the network.
The combination of these two challenges may prove to be more than what the customer is willing to tolerate. As a result, customers may stay in their comfort zone and purchase a traditional network attached storage system (NAS) and never benefit from all the advantages that object storage has to offer.
Caringo Swarm 10 — Object Storage for Everyone
Caringo provides object storage software called Swarm. Unlike most other object storage vendors, they’ve invested in providing capabilities to easily move legacy data to their object store via their FileFly software which provides automatic, policy-based data migration from Windows and NetApp NAS. They also offer SwarmNFS which provides a native NFS experience for customers but on an object store. Unlike other NFS solutions, it is designed to deliver excellent file-to-object conversion performance instead of just basic NFS compatibility.
Caringo recently released its Swarm 10 Platform and one of the significant components of the announcement is the Swarm Single Server, which provides a turnkey, standalone appliance that reduces entry-level hardware requirements by 75%. The appliance comes with 96TB raw capacity, 60TB of which is usable. It fills the entry-level unstructured data gap that many organizations face. It also enables organizations to keep data on-premises instead of in the cloud where they face perpetual storage costs, security risks, and bandwidth limitations.
Use cases for Swarm Single Server include a starter archive solution. Partnering the Single Server with FileFly, a NetApp environment could automatically offload 50TB+ of capacity from their existing filers and avoid paying for years of capacity upgrades. The solution can also serve as backup storage.
In addition to the Single Server announcement, Swarm 10 also improves performance. In a recent deployment, the core Swarm software delivered aggregate S3 throughput of 35 GB/s read and 12.5 GB/s write with no front side cache. Swarm 10 also enhances the performance of SwarmNFS, enabling a single instance to ingest 3 PB+ per month without special hardware.
However, Caringo did not leave enterprises out of the release. Larger deployments can now have clusters span across racks, rooms, and locations with minimal use of IP addressing. Caringo also simplified network administration and monitoring. Additionally, they optimized the software for dense, multi-core servers as well as adding numerous enhancements to the user interface. Lastly, they extended their Elasticsearch support to version 5, which extends the metadata search foundation and enables the power of the Elasticsearch stack and its ecosystem.
For about the past three years, Caringo has taken an “it’s on us” approach, realizing that while partnering is essential, customers seem to want a more turnkey solution. Caringo Swarm combined with SwarmNFS and FileFly broadens the object storage appeal to a wide variety of data centers. The addition of Swarm Single Server further broadens the appeal to almost every organization.