Information Technology Reference
In-Depth Information
The next step is to choose a deployment model that best suits the organizational needs.
Organizations have multiple things to take into consideration before they can arrive at a
decision. The biggest concern of them all is data security and confidentiality. Shared infra-
structure, multitenancy, and geographical location of the service provider data centers need
to be considered.
Organizations that deal with sensitive confidential customer data want to make sure the
public cloud provides secure multitenancy at every layer of the IP stack, isolation through
private volumes and LUNs, security through firewalls and secure protocols (HTTPS SSL/
TLS), and a final line of defense through data encryption. Figure 10.3 depicts multi-
tenancy in cloud storage.
FIGURE 10.3
Multitenancy in cloud storage
Users/requests
Tenant 1
Data
System
Metadata
Tenant 2
Data
System
Metadata
Tenant 3
Data
System
Metadata
Step 2: Identify bottlenecks and resolve them.
The biggest bottlenecks in most enterprise
environments are the network capacity (bandwidth), firewalls, and proxy servers. Apart
from that, policies related to data administration can also prove to be big nontechnical
hurdles. Sadly, data administration policies are completely subject to the country in which
each organization resides.
However, common bottlenecks can be resolved using common solutions. For instance,
network capacity can be increased with dedicated connections (FTP/SFTP) to the storage
cloud. The caveat is that dedicated connections are generally expensive.
To resolve bottlenecks created by firewalls and proxy servers, a number of solutions can be
devised. For instance, the firewall can be programmed to ignore certain IP addresses, thus
allowing anonymous bypass for a dedicated connection. The transfer server can be served
through a separate firewall. Caching could also be employed by the firewall to make it
faster. Figure 10.4 shows how network bottlenecks can be handled.
Search WWH ::
Custom Search