Why Flash Belongs on the Server

Brief: Why Flash Belongs on the Server Flash-based storage in enterprise IT is one of the most significant developments in the modern
datacenter. Today, almost all storage solutions supporting multiple enterprise applications utilize
flash technology. Indeed, all-flash devices are increasingly popular for certain tiers of applications.
But where does flash really belong? Should it be merely a component within an existing storage
solution, or should it be designed directly into the server architecture – closer to the applications it
Before addressing that important question, it is important to first understand the basics of
traditional storage solutions. Network attached storage (NAS) and storage area networks (SAN)
from established vendors dominate the legacy storage market scape. Storage systems from these
companies rely on storage controllers, which are not much more than x86-based servers. They are
dedicated to managing multiple shelves of disks, providing management and efficiency functions
such as snapshots, replication, compression, de-duplication, and automatically tiering data.
Each storage array commonly utilizes two storage controllers for high availability and load
balancing. Shared storage resources are connected to the server cluster through existing Ethernet
or dedicated Fibre Channel networks. With this traditional three-tier architecture, data reads and
writes travel from a virtual machine (VM) to the physical server, through the network fabric, to the
storage controller, and eventually to the disks.
How does flash fit into this model? Flash provides applications with high performance and low
latency, delivering thousands of inputs/outputs per second (IOPS), compared to just hundreds for
spinning disks. Data tiering is a common technique leveraged in today’s hybrid storage solutions,
which contain both hard disk drives and solid state drives (SSDs), to drive higher performance. The
storage controller keeps track of ‘hot data,’ or data that is most frequently used, and promotes this
data into the higher performance SSDs. When data becomes ‘cold’, it is automatically moved to
slower spinning disks.
Tiering ensures that the most commonly used data is always on the highest performance storage
media. This is very similar to how processors use DRAMs today. Since the amount of data that can
be stored in memory is limited, only the “hot data” is in the DRAM and aged data is automatically
swapped out. Why then is DRAM closer to the CPU (i.e., inside the server), whereas flash-based
storage is placed a network hop away inside a distant storage array? Does it not make sense to
keep flash closer to the server just the way DRAMs are today? After all, the PCIe bus is closer to the
CPU and provides more aggregate throughput and lower latency than Fibre Channel networking or
Ethernet over a switch.
Hyper-converged vendors, like Nutanix, who integrate compute and storage in a single 2U
appliance, have been incorporating server-side flash for the past three years. Current datacenter
architectural models are on a path to obsolescence in the new world of web-scale IT. For instance,
what will happen when there are 50 servers in the compute cluster vying to access flash-resident
data on a NAS array? SSDs capable of tens of thousands of IOPS are now connected through the
same controller on the same network. It is important to understand where the performance
bottleneck is – is it the network, the controller, or flash disks? Now, what happens if 50 or even 100
additional servers are added? Will the current model scale? It is analogous to draining water from a
barrel using a single straw. Will adding more straws solve the problem?
The problem is that flash performance is constrained by the underlying storage architecture. To
maintain performance, the end-to-end infrastructure has to scale with the demand for data.
However, anyone with datacenter experience knows that capacity planning exercises are rarely
Technical Brief: Why Flash Belongs on the Server accurate, as business and application requirements are notoriously difficult to predict. Add the
common pain points of traditional storage infrastructure, such as zoning, unmasking, LUN
provisioning, lack of VM-centric functionality, and more, and it is clear why storage administrators
struggle to maintain overall performance and availability SLAs.
Hyper-converged solutions incorporate flash disks in the server, delivering high levels of data
resiliency and pooled data storage that is equivalent to a NAS device, while allowing fast,
unfettered access to the flash storage by applications. They are designed for scalability and
parallelism, without the inherent limitations created when accessing storage through a single
controller. When deployed in clustered configurations, hyper-converged appliances avoid the
network contention that occurs when there is simultaneous demand for flash disks in a single
storage device. Instead, each server has its own flash disks, such that most read operations can be
satisfied by local, direct attached resources, and avoid network access entirely. Even for I/Os that
require the hyper-converged server to go out over the network (e.g., remote reads from adjacent
nodes, or replication of writes to other nodes), there is no single point of bottleneck. Instead, these
operations are distributed across different servers throughout the cluster.
Hyper-convergence optimizes the benefits of flash, even when cluster size grows. Further,
introducing additional storage is as easy as adding another server with the required SSD capacity
or replacing the current SSDs on the server(s) with higher capacity ones. Integrating flash into the
server enables all applications to run on common infrastructure, including workloads that are
bandwidth-hungry or sensitive to storage latency. Server-based flash also eliminates the cost and
complexity of deployment silos driven by hardware architecture.
Figure 1: Evolution of Flash from DAS to Networked Storage to Hyper-Converged Infrastructure
Flash-based storage delivers orders-of-magnitude faster performance than traditional hard disk
drives. However, inserting flash resources into central arrays and accessing them over a network
will never allow their full potential to be realized. Going forward, the emergence of non-volatile
memory express (NVMe) memory brings a lighter-weight standard for accessing SSDs, which will
only increase the demand for server-based flash. Hyper-converged architectures offer a holistic
approach to realize the potential of the modern SSD technology innovations.
About Nutanix
Nutanix is the leader in hyper-converged infrastructure, natively converging compute and
storage into a single 100% software-driven solution to drive unprecedented simplicity at
lower costs in the datacenter. Customers run any application, at any scale with predictable
performance and economics.
Learn more at www.nutanix.com or follow up on Twitter @nutanix.