Capacity planning
We measure CPU, RAM, disk and IOPS on current servers and size the hosts from that data.
Start here
Free Audits 5Urgent help
Emergency 4Ongoing support
Managed IT 5Projects
Servers & Hosting 6 Network & Virtualization 9 DevOps 5 Security & Recovery 5 Hardware & Licensing 2Find out where your servers, security, backups and performance stand, at no cost.
Server down, network out, site hacked or data lost? Call us now.
Monthly network and server support with response times written into the contract.
Setup, configuration, management and migration of Linux and Windows servers, panels and mail.
Network design and cabling, MikroTik, VoIP, branch links, virtualization and private cloud.
Networking & communications
Virtualization & cloud
Containers, Kubernetes, automated delivery, infrastructure as code and observability.
Server hardening, firewalls, backup and DR, ransomware recovery and incident forensics.
Advice, supply and installation of servers and network gear, plus genuine enterprise licenses.
Start here
Urgent help
Ongoing support
Projects
One application per physical server means more power, more rack space and hardware that sits idle most of the day. With server virtualization those services share two or three hosts, and in a cluster the VMs from a failed host restart on the others. We choose the platform around your licensing budget and workloads, whether that is VMware vSphere, Proxmox VE or Hyper-V.
Server virtualization is running many virtual servers on a small number of physical hosts, so services share hardware, power and rack space and no longer sit on idle machines of their own. It suits organizations with several underused physical servers, an out-of-warranty server behind the accounting software, or VMware licensing that became too expensive. PikoSystem measures CPU, RAM, disk and IOPS, builds ESXi with vCenter, a Proxmox VE cluster or a Hyper-V failover cluster with HA, sets up shared storage or Ceph and virtual networking, deploys VM backups with Veeam or Proxmox Backup Server, and moves existing servers across. You end up with a capacity report, cluster and storage diagrams, test restore and host failure results, admin access and a runbook for creating VMs, snapshots and restores.
We measure CPU, RAM, disk and IOPS on current servers and size the hosts from that data.
We install ESXi with vCenter, a Proxmox VE cluster or a Hyper-V failover cluster, and enable HA for critical VMs.
Depending on budget, we set up shared SAN or NAS storage over iSCSI or NFS, or Ceph on the hosts themselves, and tune RAID and caching.
Management, VM, storage and migration traffic are separated by VLAN or dedicated NICs, with vSwitches or Linux bridges configured to match.
We deploy Veeam Backup or Proxmox Backup Server, set schedules and retention, and run a test restore.
Physical servers or VMs from the old platform are moved to the new environment, and old hardware is retired only after testing.
Since Broadcom acquired VMware, perpetual licenses are no longer sold. vSphere now comes only as a subscription, bundled into offerings such as VMware vSphere Foundation and VMware Cloud Foundation. Pricing is per CPU core with a minimum core count per processor, so hosts with high-core CPUs cost noticeably more to license.
Organizations holding older perpetual licenses can keep running that version, but support and security patches depend on an active contract. That is why many teams now compare a fresh ESXi setup against moving to Proxmox or Hyper-V before renewing.
If most of your servers run Windows, Hyper-V is a natural fit. A Windows Server Datacenter license covers unlimited Windows VMs on that host, while Standard covers two, and the Hyper-V role itself costs nothing extra.
HA requires a Failover Cluster with shared storage or Storage Spaces Direct, and S2D is available only in Datacenter edition. Management runs through Windows Admin Center or System Center VMM. For Linux-heavy estates and low-cost backups, Proxmox is more flexible.
Ceph spreads storage across the hosts' own disks and removes dependence on a single NAS, but it needs at least three nodes, a dedicated 10 GbE or faster network and enterprise SSDs with power loss protection. Consumer SSDs wear out quickly under Ceph's write pattern. RAID controllers must run in HBA mode so Ceph or ZFS sees the raw disks.
Shared SAN or NAS storage over iSCSI is simpler, but a single box with one controller becomes the failure point for the whole cluster. Multipath and two storage switches matter in that design.
Snapshots are not backups. One left for weeks grows a large delta file and slows the disk. Keeping backups on the same storage as the VMs means a storage failure takes both.
Oversized VMs are common too. A VM with 16 vCPUs that uses four waits for cores to free up, and CPU Ready climbs. With HA, if no capacity is reserved for losing a host, VMs fail to restart on the remaining ones. Virtual domain controllers should also sync time from an external source.
We review current servers and services and measure resource use.
We compare platforms on licensing, features and cost with you and write the final design.
Hosts, storage and backups are brought up, and VMs are moved in agreed time windows.
We simulate a host failure, test a backup restore and hand over documentation.
VMware vSphere has mature management tooling, but its licensing costs rose after the sales model changed. Proxmox VE is open source and includes clustering, HA and backup without a mandatory license. The right choice depends on your applications, your team's skills and your budget.
For a Proxmox cluster with HA we recommend at least three nodes to keep quorum. Two servers can work if you add a lightweight QDevice as the third vote.
Yes, if the CPUs support hardware virtualization and there is enough RAM and disk. Very old servers are usually better used for backups or test environments. Capacity planning shows where each server fits.
Related searches
Quote
A few lines on where things stand and what you want is enough. An engineer calls you back, not a sales rep.