Red Hat NETWORK 3.6 - Installationsanleitung PDF herunterladen (Seite 6)

Technical white paper | HP RA for Red Hat Storage Server on HP ProLiant SL4540 Gen8 Server

The enterprise drop box or cloud storage workload can be characterized as initially comprising only writes, as end-users

upload media files. Over time reads and writes become roughly equivalent as users access, view, and play those media files.

File sizes range from a few megabytes (smart phone photos) to dozens of megabytes (HD videos taken using high-end

digital SLR cameras). The total size of the data store and its growth is directly proportional to the number of users of the

service. Deployments typically start at hundreds of terabytes and there is a need to pre-provision for potentially millions of

subscribers or enterprise users. As new users sign up for content services, typical data stores can easily grow to several

petabytes over time, and often span multiple datacenters. Data access patterns consist of hundreds of concurrent users

reading and writing at the same time.

Competitive storage solutions must be able to deliver very high sustained throughput. As shown in Figure 3, multiple HP

ProLiant SL4540 Gen8 servers are clustered via Red Hat Storage Server to provide a unified namespace, allowing access

from diverse devices in either an enterprise drop box or cloud storage setting.

Together Red Hat Storage Server running on the HP ProLiant SL4540 Gen8 server provides a scale-out architecture with

distinct advantages that include:

• Massive and linear scalability, with capacity and performance that can scale independently with no bottlenecks

• Object storage support via the OpenStack SWIFT API

• Support for multiple drop box software implementations that can be co-resident on the storage servers

• Simultaneous storage and retrieval of files and objects interchangeably

• Consistent, sustained, and predictable throughput performance

Figure 3. Red Hat Storage Server and HP ProLiant SL4540 Gen8 servers provide a unified namespace and can be deployed to serve as an

enterprise drop box, or as cloud storage for service providers.

Near-line archival

A wide variety of organizations need to backup vast amounts of unstructured data, quickly accessing archived data when

needed. Financial services firms, pharmaceutical companies, logistics operations, telecommunications, oil and gas,

healthcare, public safety, educational, and government organizations all share this common need. Complicating matters,

new regulations now dictate that organizations must comply with multi-year data retention policies. Customers are also

demanding stringent Recovery Point Objectives (RPO) and Recovery Time Objectives (RTO).

Near-line archival strategies often involve replacing tape archival with faster disk based access. Consistent access to data

regardless of where it resides with consistent performance is another key requirement. Many storage solutions are simply

unable to keep pace with the explosive growth of data and information.

The near-line archival workload can be characterized as write once, read sparingly or never. The average file sizes depend

on the kind of data that is being backed up by the organization. The total data store usually starts with dozens of terabytes

and can easily grow to several petabytes. However, if tape replacement is involved, the data store may start at a petabyte

scale. The data access pattern consists of concurrent writes as the archive is populated, and typically very few if any reads.

1 2 3 4 5 6 7 8 9 10 11 ... 28 29

Keine Kommentare

Red Hat NETWORK 3.6 - Installationsanleitung Seite 6