POWERVM - BASICS

Basics

PowerVM (Power Virtual Machine):

PowerVM formerly known as Advanced Power Virtualization, it is the virtualization solution for AIX.
PowerVM has 3 editions:

Express Edition: Hypervisor, DLPAR (3 servers), VIO (1 server), IVM, NPIV
Standard Edition: +DLPAR (254 servers), + VIO (2 servers), +HMC, +Multiple Shared Processor Pools, + Shared Storage Pools
Enterprise Edition: ++Active Memory Sharing, ++Live Partition Mobility


Intergrated Virtualization Manager (IVM)
For a smaller environment, not all functions of an HMC are required, and the deployment of additional HMC hardware may not be suitable, so IBM has developed the IVM, a hardware management solution that performs a subset of the HMC features for a single server, avoiding the need for a dedicated HMC server.
IVM manages standalone servers so a second server managed by IVM would have its own instance of the IVM. With the subset of HMC server functionality, IVM provides a solution that enables the administrator to quickly set up a system. The IVM is integrated within the Virtual I/O Server product.


POWER Hypervisor
POWER Hypervisor is the foundation of IBM PowerVM. It is a firmware layer sitting between the hosted operating systems and the server hardware, and it is always activated.
It delivers functions that enable capabilities: dedicated or micro partitioning, virtual processors, virtual ethernet- scsi- fibre channel- adapters and virtual consoles.


LPAR - Dedicated processors
Dedicated processors are whole processors that are assigned to dedicated-processor partitions (LPARs). The minimum processor allocation for an LPAR is one (1) whole processor, and can be as many as the total number of installed processors in the server.
Each processor is wholly dedicated to the LPAR. It is not possible to mix shared processors and dedicated processors in the same partition.


Micro-Partitioning
Micro-Partitioning is the ability to distribute the processing capacity of one or more physical processors among one or more logical partition.


Shared-processor pools
In POWER5-based servers, a physical shared-processor pool is a set of physical processors that are not dedicated to any logical partition. Micro-Partitioning technology coupled with the POWER Hypervisor facilitates the sharing of processing units between micro-partitions.

Multiple Shared-Processor Pools (MSPPs)is a capability supported on POWE6. This capability allows a system administrator to create a set of micro-partitions with the purpose of controlling the processor capacity that the set of micro-partitions can consume from the physical shared-processor pool. The set of micro-partitions form a unit and this can be managed for example how much processor capacity it can use.

On all Power Systems supporting Multiple Shared-Processor Pools, a default Shared-Processor Pool is always automatically defined. The default Shared-Processor Pool has a pool identifier of zero (SPP-ID = 0) The default behavior of the system, with only SPP0 defined, is the current behavior of a POWER5 server with only a physical shared-processor pool defined. Micro-partitions are created within SPP0 by default, and processor resources are shared in the same way.

If several partitions from different shared processor pools are competing for additional resources, the partitions with the highest weight will be served first. You must therefore define a partition’s weight based on the weight of partitions in other shared processor pools.


Shared Storage Pool
A shared storage pool is a pool of SAN storage devices assigned to multiple Virtual I/O Servers. It is based on a cluster of Virtual I/O Servers. When using shared storage pools, the Virtual I/O Server provides storage through logical units (file backed storage device) that are assigned to client partitions and it appears as a virtual SCSI disk in the client partition. Shared Storage Pools are using thin provisioning.


Storage Pool vs Volume Group
The IVM and HMC environments present 2 different interfaces for storage management under different names. Storage Pool interface under IVM is essentially the same as LVM under HMC. (These are used sometimes interchangeably.) So volume group can refer to both volume groups and storage pools, and logical volume can refer to both logical volumes and storage pool backing devices.


Active Memory Expansion:
Active Memory Expansion is the ability to expand the memory available to an AIX partition beyond the amount of assigned physical memory. Active Memory Expansion compresses memory pages (so it generates CPU load) to provide additional memory capacity for a partition. (It is a Power7 feature.) Starting with POWER7+,memory page compression and decompressionis offloaded to a hardware accelerator.


Active Memory Sharing:
Active Memory Sharing (AMS) enables the sharing of a pool of physical memory among partitions on a single Power server (Power 6 or later), helping to increase memory utilization and drive down system costs.


Active Memory Deduplication:
To optimize memory use, Active Memory Deduplication avoids data duplication in multiple distinct memory spaces. On traditional LPARs, multiple identical data are saved across different positions in main memory. Active Memory Deduplication combines the data in just one physical memory page and frees the other chunks with identical data. The result is multiple logical memory pages pointing to the same physical memory page, thus saving memory space. (It is available on LPARs using Active Memory Sharing.) 


Acitive Memory Mirroring:
It is called sometimes system firmware mirroring. Active Memory Mirroring for the hypervisor is designed to mirror the main memory that is used by the system firmware to ensure greater memory availability . When enabled, an uncorrectable error that results from a failure of main memory used by the system firmware will not cause a system-wide outage. The system maintains two identical copies of the system hypervisor in memory at all times.

Virtual network: 

Enables interpartition communication without assigning a physical network adapter to each partition. If the virtual network is bridged (SEA), partitions can communicate with external networks. 

Virtual switch: 

An in-memory, hypervisor implementation of a layer-2 switch.

VLAN:

A VLAN (virtual local area network) is a method to logically segment a physical network so that layer 2 connectivity is restricted to members that belong to the same VLAN. This separation is achieved by tagging Ethernet packets with their VLAN ID and then restricting delivery to members of that VLAN. The default VLAN ID for a switch port is referred to as the Port VID (PVID).

The VLAN ID can be added to an Ethernet packet either by a VLAN-aware host, or by the switch in the case of VLAN-unaware hosts. Therefore, ports on an Ethernet switch must be configured with information that indicates whether the host connected is VLAN-aware. For VLAN-unaware hosts, a port is set up as untagged and the switch tags all packets that enter through that port with the Port VLAN ID (PVID). The switch also untags all packets that exit that port before delivery to the VLAN unaware host. A port that is used to connect VLAN-unaware hosts is called an untagged port, and it can be a member of only a single VLAN identified by its PVID. 

Hosts that are VLAN-aware can insert and remove their own tags and can be members of more than one VLAN. These hosts are typically attached to ports that do not remove the tags before the packets are delivered to the host. However, it inserts the PVID tag when an untagged packet enters the port. A port allows only packets that are untagged or tagged with the tag of one of the VLANs that the port belongs to. 

Virtual Ethernet adapters:

A virtual Ethernet adapter allows client partitions to send and receive network traffic without a physical adapter. A virtual Ethernet adapter is created when you connect a partition to a virtual network. TCP/IP communications over these virtual networks are routed through the Hypervisor, by copying the network traffic (packets) directly from the memory of the sender logical partition to the receive buffers of the receiver logical partition 

Virtual Ethernet adapters allow logical partitions within the same system to communicate without having to use physical Ethernet adapters. 

Virtual Ethernet adapters are connected to an IEEE 802.1Q virtual Ethernet switch and this switch allows logical partitions within the same system to communicate with each other. You can use virtual Ethernet adapters without using the Virtual I/O Server.

Virtual network bridges:

At the Enhanced GUI on the HMC a new term is used for the SEA, which is called "Virtual Network Bridge". 

The virtual network bridge (SEA)  connects the internal network traffic to the outside world through a physical network adapter. A virtual network bridge has one or more load groups (trunk adapters). The number of load groups determines the number of virtual Ethernet adapters present on SEA.

A virtual network bridge can be associated with one untagged virtual network and up to 20 tagged virtual networks. When a virtual network is added to an existing bridge, a tagged virtual network is created. When a virtual network is added to a new bridge, it can be added as an untagged network or as a tagged network.