MegaFon’s driving principles are fair competition and maximum subscriber satisfaction. In order to achieve these goals, they strive for an efficient and smooth-running corporate IT infrastructure.
The MegaFon IT group focuses on a number of “must-haves”:
“Our decision to virtualize was more than just a financial matter,” says Ivan Iotchenko, Systems Management Group Engineer at MegaFon. “We could not guarantee flawless work for some of our critical applications - means to ensure the reliability of these applications was either not provided by the manufacturers, or the price of required functionality was too high - and that was the problem.”
MegaFon chose VMware virtualization technology to create a robust and cost-effective IT infrastructure. The initial deployment included several VMware ESX 3.5 servers arranged in a DRS cluster. The hardware consisted of 4-processor HP ML570 G2 and HP ML570 G3 servers with a SAN based on HP MSA2212fc.
By early 2009, MegaFon had virtualized close to 40 servers, all managed by VMware vCenter, which is running on an HP DL360G4p server. The company is steadily growing its virtual environment – as stated by MegaFon’s experts, planned virtualization growth is 10-15 new virtual machines per quarter with a target of up to 100 virtualized servers by 2010. As the underlying hardware for these new virtualized servers, MegaFon is planning to use HP blade servers.
So far, MegaFon has virtualized production applications such as SMS services, web servers processing subscribers’ requests, and management systems for production equipment. “In addition to that, we virtualized monitoring services that were missioncritical to us. We decided to move them onto our VMware virtual infrastructure because of the high availability that the ESX cluster provides,” explains Ivan Iotchenko.
Having moved VMware vSphere into production, MegaFon next faced the problem of centralized monitoring of IT services. The company uses the enterprise class HP Operations Manager for Unix to monitor the performance and health of physical servers, using HP agents deployed on the physical servers. MegaFon IT operators use the HP OM console as their primary management interface.
The problem is that HP agents installed in virtual machines, lacking visibility of the underlying physical infrastructure, cannot provide an integrated, accurate picture of the virtual environment. Critical VMware virtual infrastructure metrics, such as vCenter events, CPU ready time, balloon memory, and memory swap rate, are simply not available to HP agents running in a virtual machine.
“When we put VMware in place, we realized that we needed to gain insight into how our virtual infrastructure was performing,” says Ivan Iotchenko. “We did not want to buy, deploy and maintain another standalone monitoring solution – we wanted to have data from our VMware virtual infrastructure incorporated in HP Operations Manager. We needed a product that would leverage the capabilities of HP Operations Manager and help consolidate data from both physical and virtual servers across the whole enterprise in a single monitoring console.”
In March 2009, the Complete Company, a Veeam Gold ProPartner in the Russian Federation, offered MegaFon the nworks Smart Plug-in for VMware – an enterpriseclass solution for monitoring the performance and health status of the entire VMware infrastructure in HP Operations Manager.
The nworks solution fully integrates with HP Operations Manager without the requirement for HP agents running in the ESX host console operating system, which is critically important for the stable performance of the IT infrastructure. “Agents always mean additional workload on the server. When we talk about mission-critical applications, agents become an area of potential risk,” explains Ivan Iotchenko. Based on VMware’s strategic VI API and in contrast to alternative solutions, the Veeam nworks Smart Plug-in for VMware uses the VI API to collect VMware-specific metrics and events, reducing the monitoring impact on ESX hosts and vCenter servers.
The nworks Smart Plug-in for VMware includes a set of out-of-the-box thresholds, targeting, event suppressing for monitoring events, metrics of VMware virtual infrastructure, and the Knowledge Database providing detailed technical information about VI metrics, alerts of different severity levels and so on. ”There is hardly a person who knows all the nuances of VMware VI,” continues Ivan Iotchenko. “Preset monitoring policies and the Knowledge Base speeded deployment of the nworks solution and increased work efficiency of our monitoring team. Once you install it - it’s ready to use.”
The nworks Smart Plug-in for VMware was highly praised by MegaFon experts owing to:
VMware virtual infrastructure yielded tremendous benefits to MegaFon: effective use of hardware resources, reduction of hardware and maintenance costs, fast services deployment, high availability and fault tolerance of IT systems. “New technologies always bring new challenges; in our case, centralized monitoring of IT systems was that challenge,” concludes Ivan Iotchenko. “With the nworks Smart Plug-in for VMware, we can take advantage of the benefits associated with virtualization and HP Operations Manager. Rather than spending extra time and resources on deployment and maintenance of another standalone monitoring solution for VMware infrastructure, we can focus on providing better services to our customers.”