IT automation tools impose serious software and hardware requirements, which change depending on your data center and the number of devices under automation.
Automated IT tasks require real-time operations across thousands — maybe tens of thousands — of devices. Inadequate computing resources impair the automation tool’s performance and prevent it from scaling. Properly installing the right automation tool might require hardware upgrades or other architectural changes.
There are a few points you should follow when sizing up an automation tool. First, know how many devices it will automate — now and in the future. Next, size the tool’s server to offer computing resources that accommodate the maximum anticipated device load into the foreseeable future (at least until the next server technology upgrade). Finally, test the automation tool’s deployment thoroughly in a proof-of-principle project to verify performance under load.
Hardware requirements
Many tool deployments run into trouble because the project uses servers with inadequate computing resources for the number of devices being managed.
Consider the suggested requirements for Kaseya Server 6.3. The minimum server configuration is a single 2.4-GHz processor core, 4 GB of RAM, a 40-GB hard drive and a 100-Mbps network port. Most production servers today far exceed these minimum requirements. But that’s the minimum configuration. The tool will launch and run, but likely present performance problems when managing more than a few systems.
Automating tasks on up to 1,000 machines takes a dual-core 1.86-GHz processor with 10 GB of RAM, three 73-GB 10,000-rpm Serial-Attached SCSI (SAS) drives in a RAID 5 configuration and a 1-GB Ethernet port (or similar reservations if deploying the tool in a virtual machine [VM]).
To support 5,000 machines, the server requirements jump to a quad-core 2.26-GHz processor with hyper-threading (eight cores) or a dual 2-GHz quad-core system, 18 GB of RAM, four 73-GB 15,000-rpm SAS drives in a RAID 10 configuration and a 1 Gigabit Ethernet (GbE) port.
By the time your tool is automating 10,000 systems, requirements include dual quad-core processors with hyper-threading (16 cores), 30 GB of RAM, eight 73-GB 15,000-rpm SAS drives in RAID 10 for the database, two 146-GB 15,000-rpm SAS drives in RAID 1 for the operating system and a 1 GbE network interface card with multiple ports. A single modern server can still accommodate this computing load, but note how the requirements jump as the deployment scales up.
With agents, software installs on each managed system and allows the automation tool to receive data and exert control over an array of heterogeneous devices across the enterprise. This means agent software management — patches, updates — on each system.
And since the agent is executed on each managed device, the agent can increase the computing load on each device. The good news is that the computing overhead for typical agents is light and usually works well on any system hardware that accommodates the supported OS. For example, the agent for Kaseya Server 6.3 requires a 333-MHz processor, 128 MB of RAM and 30 MB of disk space. As with the automation tool itself, it’s sound practice to test agent behavior and benchmark major systems to quantify the actual performance impact.
Software requirements
The underlying OS is also important to the IT automation tool. For example, comprehensive frameworks like Kaseya Server 6.3 operate on Microsoft Windows Server 2003, 2003 R2, 2008, 2008 R2 or 2012. Evaluations might be possible on endpoint OSes like Windows XP Pro, Vista, 7 and 8, but a server OS is essential for production deployment.
By comparison, Puppet Enterprise and similar tools are primarily for Linux servers running Red Hat Enterprise Linux 5 and 6, CentOS 5 and 6, or Ubuntu 10.04 and 12.04, among others. AIX, Solaris and Windows Server platforms are only supported with agents. Select your IT automation tool with your predominant data center OS in mind.
Automation tools organize an incredible amount of data, so database software is essential. Expect to use a database like Microsoft SQL Server 2005, 2008, 2008 R2 or 2012 with the latest service pack. The tool might need SQL Reporting Services to organize and extract the data collected in the database. Puppet Enterprise includes a database component.
Additional software requirements vary. For example, Kaseya Server 6.3 requires Microsoft Internet Information Server 5.1 or greater for Web-based access; Microsoft Message Queuing allows applications running at different times to exchange messages across heterogeneous networks and systems (even offline) and the Microsoft .NET Framework 2.0 to 3.5 SP1.
In virtualized data centers, the tool must recognize VMs to monitor, deploy and manage tasks. Will the tool work as well running as a VM as it does in a physical server? The challenge is to reserve adequate VM resources for the automation tool.
Source: http://searchdatacenter.techtarget.com/tip/Sizing-up-hardware-resources-to-allocate-to-IT-automation-tools?asrc=EM_NLT_24117347&utm_medium=EM&utm_source=NLT&utm_campaign=20131014_Does%20Unix%20still%20have%20a%20role%20in%20data%20centers?_ewatkins&track=NL-1811&ad=889373