Page 77 - Cloud Essentials

be anywhere (especially public cloud clients), bandwidth

considerations are paramount.
Latency tolerance Cloud computing serves clients at spread-out
locations, and you should expect an increase in transmission delays.
Not all applications tolerate these delays.

Firewalls and access control This can be configured in the cloud,
locally on premises, or both. For example, using IaaS cloud servers
and networks implies the use of a private cloud network. Some

providers allow you to use your existing IP address structure in the
cloud. This private cloud network might have configurable firewall
options. Cloud servers (virtual servers) always have their own
firewall settings.
Resilience and redundancy Virtual servers, clustering, load

balancing, cloud storage, backup, data replication—these all ensure
that computing services are highly available and that there are multiple
copies of the data they produce.

Some or all of these points will be detailed with your cloud provider–
supplied service level agreement (SLA). The SLA spells out what cloud
customers can expect in terms of service including guaranteed uptime,

response time, and billing rates.

Automation and Self-Service
Traditionally, IT systems are built with engineering diagrams and work

orders, which are then executed by IT operations, and this can take weeks.
Cloud-computing service models allow that to be done through self-service
web sites and web services, which can take merely minutes.
Automation is the key to reducing operational expenses. How automation
is implemented depends on the existing capabilities of the organization. The
desirable end state is where business requirements are input to a service
orchestration tool or function that automatically configures entire

application server stacks and automatically provisions and deprovisions
servers, as load varies. According to the National Institute of Standards and
Technology (NIST), “Service Orchestration refers to the composition of

77