We selected a few products that sell at a comparable price point. These products also run CIFS/SMB, so they have a strategic advantage over the B810i. The QNAP TS-831X has the added benefit of onboard 10-Gigabit Ethernet from the factory, and you can optionally purchase additional hardware to enable it on the IX Systems FreeNAS XL. The Asustor AS6210T doesn't include provisions for 10GbE because it lacks a PCI Express port to add an Ethernet card.
Our test system is a modified Quanta MESOS CB220 Cluster-In-A-Box multi-node server.
We upgraded the system with two Intel E5-2670 V1 processors and 128GB of system memory per node. An Intel X520 network interface card with two 10GbE ports resides in each node. The system allows us to deploy four test nodes that execute custom software scripts using industry standard test software to measure our base performance metrics.
The test system and devices under test connect via a Netgear M6100 fully-modular switch, which provides ample 10GbE and 1GbE ports to sustain our large testing environment.
Drobo B810i Setup
We used the Drobo setup application to build our iSCSI storage, just as you would during a typical deployment. We loaded the system with eight Seagate Enterprise NAS HDDs for the first round. We removed two HDDs and replaced them with SanDisk Extreme PRO 240GB SSDs to test the cache features.
We ran a single gigabit Ethernet line to each system to simplify the setup and to provide a level playing field. Some of the systems support two or more Gigabit connections, and in some cases, 10-Gigabit.
We began our testing with simple synthetic workloads to gauge baseline performance. The synthetic tests allow us to examine performance at various block sizes and queue depths. After looking at the Drobo 5N and B810n specifications, most readers shouldn't be surprised to see the B810i at a severe performance disadvantage compared to the systems shipping out of Taiwan. The B810i uses an older Marvell SoC that is long past its prime.
The Drobo Data-Aware Tiering feature doesn't do much for our synthetic workloads. We did notice a nice improvement with trace-based testing that you will see later in this review.
Like the Drobo B810n, the B810i running over iSCSI doesn't provide much of a performance improvement as we ramp up the queue depths.
Mixed Sequential Workloads
Mixing reads and writes doesn't seem to hurt performance as much on the B810i as it does on other systems. The performance line shoots across the charts with minor variation.
The Drobo B810i delivers very good random read performance.
We start to spot some performance improvements with the automated data-aware tiering during the random write tests. Even with the cache, the B810i fails to meet the same level of performance as the QNAP TS-831X with a Gigabit connection. The TS-831X retails for the same price as the X810i, includes 10-Gigabit onboard, supports CIFS, and can also use an SSD to cache I/O.
Mixed Random Workloads
Mixed random workloads show that the SSD cache will benefit users even when a small amount of the workload includes random writes.
Standard Server Workloads
The server workloads reveal that the Drobo B810i can outperform the other systems during light workloads. The SSD cache didn't have much of an impact on many of the tests. However, the SSD cache benefited the email workload and provided lower latency.
iSCSI Consumer Workloads
With iSCSI, you get a drive letter that you can use to install software, just like a local disk. That opens the door for diverse types of software running simultaneously on the SAN. The Drobo B810i struggles through most of the tests and consistently falls to the tail end of the chart. The SSD cache-enabled results paint a much brighter picture for the B810i, but that may be short lived. With two SSDs, the Drobo B810i can match the other systems operating without an SSD cache. The problem is the other systems can also use SSDs to deliver enhanced performance.
Total Storage Bandwidth
The Drobo B810i falls far below the other products that sell for around the same price, but they have more complicated setup procedures. We also see how the data-aware tiering improves performance with real-world applications. The cache doesn't respond well to our synthetic testing, but Drobo claims it can double performance with a real workload.
MORE: Best SSDs
MORE: All SSD Content