Banner
Backup D.R. Replication Virtualisation Hardware/Media Privacy

Current Filter: Storage>>>>>>

PREVIOUS

   Current Article ID:2816

NEXT



Tiered Storage: at the Heart of the Data Centre

Editorial Type: Opinion     Date: 09-2013    Views: 3061   






Nick Spittle, Director of Toshiba Electronics Europe, Storage Products Division examines how a tiered storage architecture can exploit the best features of both SSD and HDD, allocating data according to access needs

The world of storage is changing rapidly; data is being created and shared at a phenomenal rate. With increasing internet speeds, workplace reliance on IT systems, access to HD and 3D video content and the rise of social media, digital data production and storage is at an all-time high - and businesses are at risk of being overloaded by data. Market analysts IDC estimate that the amount of data created or replicated in 2011 was 1.8 zettabytes, the equivalent of 57.5 billion 32GB iPads filled with 200 billion high definition movies.

Businesses are increasingly asking how to meet the ever-increasing storage demands in a cost-effective manner that meets all performance needs such as data availability, speed of access, security and data redundancy. The majority of digital data is still stored on HDDs, the foundation of almost every enterprise data centre. However SSDs offer compelling advantages such as lower energy consumption and faster data access times. The solid state drive (SSD) has been hailed as the future of the storage market but is it truly the right technology to meet the digital avalanche that is coming our way?

PERFORMANCE
HDD performance will always be at the mercy of the mechanical components used in the drives. These components need continuous correction in order to maintain position and speed of movement. HDDs combine rotating magnetic discs with read/write heads held in position just above the surface by an arm with a positioning mechanism. Data is written and read by generating or sensing a magnetic field and the data is held and accessed within a sector/track and cylinder construction.

Because the read/write heads have to move into position and the disc needs to 'spin up' to speed, there are inevitable delays (latencies) in data writing and retrieval. Drive access times are slowed by command processing, seek times, rotational delays and data transfer times. Conversely, SSDs have no moving mechanical element to their design, removing any seek time and rotational delays. SSDs store data on NAND flash memory, which is accessed via a chip/block/page layout and cluster/sector construct, where under eSSD architecture a cluster is the minimum readable unit (8 sectors).

The boost in data flow rate and input/output per second (IOPS), is one of the major selling points of this technology. However, SSDs face their own limitations in terms of life expectancy. Workload, especially in the enterprise sector can be extreme and involve high volumes of transactional data storage, which could be weighted towards write operations over read. For SSDs the frequency of this data change predicts the life of the device in the field.

COST
Price is the primary factor where HDDs win over SSDs. HDD cost per GB is lower than that of SSD, which is why HDDs are more prevalent in enterprise storage centres. However, if absolute cost is considered - especially at lower capacity points - SSDs often offer an advantage.

SSDs are still predominantly limited to high-end devices due to the fact that they're more expensive but as NAND prices fall and device sizes shrink, this is changing. Of course, device cost does not incorporate the total cost of ownership and usage, where the increased speeds and efficiency of SSDs begin to make the argument more favourable.

CAPACITY
HDDs have maintained a significant capacity advantage compared to SSDs. While enterprise HDDs with space to store more than 4TB data are available, SSDs tend to have lower capacities. Toshiba has recently announced the availability of eSSDs that have capacities up to 1.6TB and enable data transfer rates of up to 12 Gbit/s as well as highly efficient error correction codes (ECC). However, HDDs have a higher storage density and can store more data per unit volume than an SSD. Assuming the same form factor, it would take five 1.6TB eSSDs to replace just two 4TB HDDs - and with organisations often having limited data centre space the increased storage density makes a compelling case for HDDs.

ENERGY CONSUMPTION
SSD's do not have any mechanical moving parts, run cooler and consume less energy than an equivalent HDD. In an enterprise situation, this can be a distinct advantage where IOPS per watt is a priority as it results in lower energy costs both for reading and writing data, and data centre cooling.

LIFETIME
Unlike magnetic disc platters, NAND flash cells have a finite life expectancy, and the erase/write process erodes that life over time. The typical lifespan of an SSD depends on the type of NAND Flash used, Typically, the life span of memory cells for multi-level cells (MLCs) used in consumer SSDs is 10,000 cycles, and single-level cells (SLC) used in eSSDs is 100,000 cycles.



Page   1  2

Like this article? Click here to get the Newsletter and Magazine Free!

Email The Editor!         OR         Forward ArticleGo Top


PREVIOUS

                    


NEXT