Monitoring Cloud Infrastructure Performance to Eliminate Visibility Gaps
The ongoing adoption and consumption of cloud services presents performance monitoring challenges for IT departments. Whether you are automating the provisioning of virtual machines, networks, and storage in a private cloud, migrating your enterprise applications to a public cloud, or leveraging the growing set of Software, Platform, or Infrastructure as a Service models, your performance monitoring platform should support your journey to the cloud.
Private clouds afford you direct access to monitoring the compute, network, and storage elements. But Software as a Service (SaaS), Infrastructure as a Service (IaaS), and Platform as a Service (PaaS) models present unique challenges. With these models, you lose direct control of – and visibility over – the assets comprising the services. How do you monitor the performance of services you don’t own, such as SalesForce.com or public cloud offerings for offloading development/test workloads?
In addition to cloud service availability and optimization, you must also measure consumption levels and monitor the impact of cloud service access on the rest of your infrastructure.
Cloud service models offer customers varying levels of control over assets and services, which presents performance visibility challenges.
This document reviews the common characteristics of cloud computing and reveals how SevOne ensures performance visibility for each service model.
Did You Know?
Gartner predicts that the bulk of new IT spending by 2016 will be for cloud computing platforms and applications with nearly half of large enterprises having cloud deployments by the end of 2017. (Source: Gartner Says Cloud Computing Will Become the Bulk of New IT Spend by 2016)
Essential Characteristics of Cloud Computing
The National Institute of Standards and Technology defines five cloud computing characteristics:
|Cloud Computing Characteristics||How SevOne Helps|
|Broad Network Access
A variety of devices and locations typically access cloud resources. Providing performance visibility for the range of cloud environments requires a performance management platform capable of integrating a variety of data types across a range of infrastructures. Traditional SNMP polling techniques alone can leave performance visibility gaps.
|SevOne recognizes the variability of cloud deployment models requires a performance monitoring platform capable of acquiring and correlating a broad set of data. SevOne extends performance visibility by leveraging multiple methods for data collection and analysis, such as SNMP, NetFlow, IP SLA, and IaaS/PaaS-based APIs to understand response times and health of the network and system infrastructure required to deliver your cloud service. In addition, SevOne provides WiFi monitoring so you can see how cloud services drive demand for access across a varying set of end user devices (mobile phones, tablets, laptops, etc.).|
Clouds pool network, compute, and storage resources shared across a range of users, including multi-tenant environments. Enabling performance visibility for multi-tenant cloud environments requires tenant-by-tenant isolation, with careful attention to user and data access permissions. As an enterprise supporting multiple business units, you will likely want options to share the network, compute, and storage resources across each of your & "customers" in a multi-tenant environment.
|SevOne enables you to define a performance visibility strategy with the flexibility to report segregated experiences for different end-users in a controlled manner, securing visibility of relevant devices and objects. Even if you are not supporting multiple business units or multiple customers today, SevOne prepares you to make that transition if your management team asks you
Choosing a performance visibility strategy with possible multi-tenant use cases in mind will allow your team to define policies that automatically group and classify discovered devices, and ensure restricted access via authentication.
A consumer can unilaterally provision computing capabilities, such as server time and network storage, as needed automatically without requiring human interaction with each service provider.
|When users are able to activate cloud services in a self-service manner, they may inadvertently create increased traffic load on your network.
SevOne provides real-time, intelligent alerts when any metric in your environment deviates from normal behavior, like an abnormal spike in bandwidth consumption. In addition, single click metrics-to-flow analysis reveals the composition of traffic, so you can easily see offending users and sites.
Clouds typically leverage self-service portals with automation scripts for the rapid turning up and down of compute, network, and storage services.
|SevOne utilizes robust APIs designed to interface with compute, network, and service management automation systems for notification of the full life cycle of new device/service creation, migration, updates, and retirement. Through integration with a customer's chosen automation system, SevOne provides performance visibility for these rapidly changing environments.|
Cloud systems automatically control and optimize resource use by leveraging a metering capability at some level of abstraction. Resource usage can be monitored, controlled, and reported, providing transparency for both the provider and consumer of the utilized service.
|SevOne provides a means to measure your cloud service consumption and challenge billing claims from vendors who provide usage-based services. In addition, SevOne provides real-time views of the impact of cloud service consumption on the rest of your infrastructure. NetFlow allows you to understand the makeup of traffic traversing your network.|
Monitoring Private Cloud Environments
In private cloud deployments, IT operations teams typically have direct access to the compute, network, and storage objects comprising their cloud offering, but the introduction of automation systems for rapid provisioning of these environments can create performance visibility challenges.
SevOne helps by integrating directly with the cloud provisioning systems. When new devices and systems are added by cloud systems, SevOne can immediately monitor these new objects in your private cloud, providing you with visibility into both physical and virtual elements.
SevOne monitors physical and virtual compute environments from a single pane of glass, analyzing host and guest consumption of CPU, memory, disk, and throughput. SevOne automatically baselines this data, so in case of a memory leak or disk capacity shortage, you receive a proactive alert to help avoid potential disruptions to the business.
In the VMware virtual server space, SevOne uses the VMware vSphere API to dynamically discover the ESX hosts associated with a vCenter. An administrative user can view and select the host devices and their guest VM objects for monitoring. SevOne automatically discovers new hosts as they are added to a vCenter, collects their key performance indicators, and baselines their performance.
SevOne provides visibility into virtual server environments
In addition, SevOne supports the collection of Java Enterprise Edition (JEE) application metrics via the Java Management Extensions (JMX) plugin and Windows server application metrics from Windows Management Instrumentation (WMI) plugin, both included with SevOne out of the box.
performance indicators, you can monitor VMware, Java, and Windows server and application environments in their private cloud from the same platform you use to monitor your entire network. Benefits include being able to automatically avoid potential outages, and proactively identifying problem infrastructure and applications before end-users call the help desk.
Monitoring a SaaS Public Cloud Offering
Many enterprises are accustomed to defining performance visibility strategies for devices, services, and applications under their direct control. The shift to "as a service" models - especially SaaS models where network operations teams do not manage or control the network, servers, storage, operating systems of the service - creates a performance visibility challenge for most organizations.
For a SaaS offering like Salesforce.com, SevOne helps by leveraging of series of embedded network and system instrumentation along with the automated baselining and reporting of the SevOne platform to effectively monitoring the performance of a SaaS-based implementation.
- SaaS Performance Overview Using IPSLA
Leveraging the embedded IPSLA instrumentation in the network, SevOne collects and reports on the key connection components, including page load, TCP and DNS connection time. Dashboards can be created showing general health and notifications, with both headquarters and remote office response times to Salesforce.com, along with an internet response time comparison to a user-defined set of leading web sites.
- SaaS Performance - Leveraging NetFlow Analysis
SevOne dashboards leverage NetFlow collection, analysis and reporting to show you who within your organization is utilizing a SaaS application the most and in what direction their traffic is flowing.
- SaaS - Link Performance with Baselines and Time over Time
Understanding the primary and secondary link performance to and from your main sites to the SaaS offering rounds out SaaS performance visibility. SevOne automatically baselines each of these data points, and will raise an alert condition when thresholds are crossed, while also depicting a view of "yesterday" vs "today" of response time for the monitored links.
For SaaS offerings such as SalesForce.com, SevOne compares response time metrics (including DNS, Connections, and Load Time) across multiple sites.
Monitoring IaaS and PaaS Environments
In IaaS and PaaS deployments, IT operations teams do not manage the underlying cloud infrastructure, but have control over the operating systems and applications deployed in IaaS, while in the PaaS model, cloud providers limit controls to the application.
IT operations teams can leverage similar performance visibility strategies as SaaS models, but there's more that can be done.
SevOne can help.
Trends are emerging for IaaS/PaaS providers to not only offer portals into the operational health of the underlying infrastructure, but also offer APIs to the data that feed these portals. SevOne users can now leverage SevOne professional services to provide a more complete picture of the IaaS/PaaS environment from providers such as Amazon Web Services (AWS)
SevOne users can leverage the open APIs of the SevOne platform for performance visibility of AWS instances
AWS' management portal, CloudWatch, provides free API access to 5 minute polling interval data for a range of EC2 region, instance and volume performance data including, but not limited to, CPU utilization, disk reads/writes, network in and out, volume read/write, volume total read/write time, volume idle length and volume queue length. For an additional fee, users can obtain 1 minute polling interval data. CloudWatch also provides a set of system status check for attributes such as loss of network connectivity and loss of system power, while notifying users of per instance status checks such as exhausted memory and corrupted file systems.