Diagnosis and Performance Monitoring - Deploying and Managing a Cloud Infrastructure - page 137

between a data packet and its destination, the hop count would be exactly three counts. Of
course, this is not as simple in today's complex web of networks. Back in the days of enter-
prise data centers and networks, the hop count could be determined easily. Now that the
Internet is made up of varied networks owned by different entities with different settings
and connections, the number of hops is not so easy to count.
The hop count is related to how many network nodes (like routers) the data goes through to
reach its destination, but the shortest hop does not often correspond to a geographical shortest
path. For example, the shortest path from London to Paris might go through Quebec, which
does not make sense geographically but because of how networks on the Internet crisscross, it
might just make sense in the number of hops.
A hop count can be measured through various networking tools and applications, with
traceroute being the most common example.
Quality of Service (QoS)
Even with the cloud described with terms like homogeneity and simplicity of service , the
truth is that everything relating to the cloud is one hot, piping jumble of complexity, maybe
not in theory and concept but certainly in application and practice. The illustrations and
explanations all look so simple, but the truth of the matter is that the age-old adage of
“Better said than done” is quite applicable here. We have all the complexities coming from
the extensibility and diversity of networks within the Internet, from the complexity of distrib-
uted computing and virtualization, and of course from having to please all of our customers
who have zero tolerance for any sort of downtime or latency. Companies are pushed to create
better Quality of Service (QoS) models or simply get creative creating their guarantees and
service-level agreements (SLAs), getting “lawyery” for lack of a better term.
As part of the QoS promise, SLAs lay out what the service provider is offering the
customer and what they are prepared to do in case they cannot deliver on this promise. But
traditional SLAs do not usually differentiate between kinds of outages such as on a server,
a NIC, an NFS, or even a security vulnerability or exploit. So they might be able to get
away with a storage facility outage not being considered downtime because the rest of the
service is still accessible.
But in a nutshell, QoS is the ability of a service provider to prioritize certain applications,
users, and data flows and guarantee a certain level of performance in its applications and
services. There are numerous QoS criteria, and they may vary depending on the application
in question and include performance criteria like throughput, latency, delay, and others.
Multipathing
Multipathing is a technique that enables usage of more than one physical path to transfer
data between a host and an external storage device. This is shown in Figure 5.5. In case of
a failure of any element in the network—such as an adapter, switch, or cable—the network
can switch to another physical path and avoid the failed component. This ensures uninter-
ruptable, correct, and complete transfers with minimum failures.