At the heart of legacy IT, hosting, managed service providers (MSP) and clouds are the common building blocks that include processing, networking and, of course, storage, all of which enable and advance cloud capabilities. In addition, as one of three primary building blocks of the cloud computing infrastructure, storage has a multi-tenant role.
This article explores tiered storage media or devices for public and private clouds, including the best storage options for different cloud services based on application performance, availability, space capacity, economics and functionality requirements. Tiered storage combines solid state drives (SSDs) and hard disc drives (HDDs).
Background and Challenges
As cloud computing and cloud storage adoption continues to grow, so do the challenges of supporting access in a safe, cost-effective way while meeting quality of service (QoS), service level agreements (SLAs) and service level objectives (SLOs). Cloud computing and cloud storage on a public, private and hybrid service address information demands and associated technology changes.
People and information are both living longer and, as a result, the associated storage required to protect, preserve and serve their needs is changing the way we think about data storage. In addition to needing cost-effective data protection and preservation, performance or responsiveness of cloud-based information services and applications are critical factors in scaling cloud storage. Other business and technology challenges include being able to physically manage and store more data, along with associated processing and networking resources, in a given amount of space or footprint. The challenge is that not all applications or information services have the same storage performance, availability, reliability, space capacity, density or cost requirements.
The Need for Tiered Storage
From a cost perspective, using only high-capacity, low-performance storage, such as nearline HDDs, could possibly meet financial goals. The challenge with this, however, is that some applications would end up experiencing performance bottlenecks. On the other hand, using ultra-fast NAND flash SSDs would result in an exceptionally responsive system, but the cost to support large-scale amounts of data, using only SSDs, would simply be prohibitive.
The best solution is to use different tiers of storage devices, installed into cloud computing servers and storage systems, to meet your specific requirements. Different tiers of data storage are used in various ways by cloud services and product providers to meet diverse application needs. Your specific return on innovation for storage to support cloud applications will be measured by your system’s performance, availability, capacity and energy efficiency and how they meet various SLOs and SLAs.
The historical information usage and access model has traditionally included data creation, content modification, ending the productive lifecycle quickly, then living on a storage platform as legacy content. Once data is inactive, it might not be used again for a long time, if at all. This legacy trend still exists today, particularly for transactional and little data, making the case for data footprint reduction (DFR) techniques, including archiving, compression, deduplication, thin provisioning and storage tiering.
The challenge with legacy content is that this inactive data is occupying storage space capacity that could otherwise be used for active data. The idea is to move inactive data from premium higher-performance storage media tiers, such as SSD or fast 15K HDDs, to slower, larger-capacity, energy-efficient and lower-cost HDDs. By removing inactive data from more expensive, high-performance storage, those resources are then used more effectively while reducing overall storage costs and thereby improving the return on investment. Inactive data can be moved to a nearline or offline tier of local or offsite storage.
With a traditional data lifecycle model, there is also the notion that data or information loses value over time. This may have been the case in the past, but not now.
The new data lifecycle model shows how data can maintain or even increase its value over time by being online or nearline accessible. Many cloud-based applications, services and functionality fall into the activity model where data is created, then read, copied, analysed, downloaded and new versions created. Recognising activity patterns is an important step in aligning the approached tier of storage medium to different applications for public and private clouds.
What are Your Options?
The specific type or tier of storage for enabling your cloud capabilities will depend on the type and classes of service being provided. For example, if the focus of a cloud service is to provide low- or no-cost capabilities that do not need very high performance, then high-capacity, energy-efficient HDDs are an option. On the other hand, if the service requires fast response time (latency), while supporting a large number of active users, then SSDs are a good fit. What this means is that different tiers of storage media are aligned and used to exploit the various attributes of those technologies to changing needs.
Table 1. Various Data Storage Tiers
||SFF and LFF SAS SSD
||SFF Fast SAS 10K/15K HDD
||High-capacity SAS HDD
||SAS and SATA Tape
||Online look-up tables, indices, databases, log or journal files, VM/VDI files
||Online active files, database tables, VM/VDI files, web and user files
||Nearline: online archive videos, photos, D2D backup, snapshots
||Offline: master backup, DR gold copy, archive, long-term retention
||Very high I/O activity, low latency, I/O consolidation
||Moderate I/O activity and capacity in dense footprint
||Occasional read access, high capacity, low cost
||Infrequent access, high capacity, very low cost
||Higher I/O activity vs. space capacity, mix of reads and writes
||Higher ratio of storage capacity to I/O activity, reads and writes, random and sequential
||Focus on high storage capacity at low cost with fast random access to data
||Focus on high storage capacity at low cost for bulk access to data
||Pulsar® SAS SSDs
||Savvio® 15K and 10K 2.5 inch (SFF) SAS HDDs
||Constellation® 3.5 inch (LFF) SAS HDDs
||Various tape and virtual tape solutions
Seagate and Cloud Storage
Seagate is the leader in enterprise storage and, not surprisingly, is also at the heart of the cloud infrastructure. Exploiting decades of experience in high-density, large-scale enterprise, institution, government, managed services and co-location environments, Seagate extends that knowledge to public and private cloud environments. In addition to industry-leading storage technology, Seagate has decades of experience working with various partners and their respective storage solutions, packaging, chassis and enclosures, testing and verification processes.
As a key supplier to private and public cloud and managed services providers, Seagate technology is found in enterprise environments and cloud data centres, and is used by managed service providers to small businesses and consumers. In other words, Seagate has been enabling cloud computing and cloud storage from the data centre to your pocket for some time!
Storage options for cloud computing and cloud storage environments from Seagate include the Pulsar® family of ultra-high performance SSDs. Complementing Pulsar drives are high–performance 2.5 inch Savvio® 10K and Savvio 15K HDDs for higher-density scenarios, as well as energy-efficient Constellation® HDDs that can support a multiple terabyte configuration.
Table 2 shows how and where Seagate enables public and private cloud computing and cloud storage. Additionally, Seagate works with various partners to help produce public cloud, MSP, co-location and private cloud solutions.
Table 2. How Seagate enables the Cloud
|Data Centre: Public, Private, Hybrid
|Mix of high performance and capacity
||Cost-effective, high capacity, energy saving
||Local and cloud storage
||Local and cloud storage
|Pulsar® (SSD), Savvio®15K and Savvio 10K 2.5 inch performance-optimised drives
||Constellation® and Constellation ES capacity-optimised drives
||GoFlex® Home network storage, Backup Plus personal storage and Satellite™ mobile wireless storage
Summary and Next Steps
Many types of cloud service and product are available to support application and information needs. If you are designing or building your own cloud service or private infrastructure, and deciding what type of product to use, some decisions around resources must to be made, and made to a given cost. Tiered data storage meets the needs of applications or information services when performance, availability, space capacity, security and economic characteristics are important. Seagate understands and can exploit knowledge of enterprise IT and high-density computing environment data storage for cloud environments or decision making.
Learn more about cloud computing and cloud storage topics in the Solution Centre.