v Sphere Deepdive Magnus Bergman Joel Lindberg Agenda

v. Sphere Deepdive Magnus Bergman Joel Lindberg

Agenda § VMware v. Cloud® Suites Launch Context and Product Set § v. Sphere 5. 0 Recap § v. Sphere 5. 1 Overview • • Compute, Storage, Network—Enhancements and Features Availability, Security, Automation—Enhancements and Features v. Center Server—Enhancements and Features Additional Features and Enhancements— “The Best of the Rest” § Memory, CPU and Network Best Practises 2

VMware v. Cloud Suite 3

VMware v. Sphere 5. 0 v. Center Server New HA Architecture • v. Motion over higher latency links • ESXi Firewall • Virtual Appliance • Web Client • 32 way SMP • 1 TB VMs Application Services Availability Security Scalability Compute Storage Network VMware v. Sphere 5 Infrastructure Services • ESXi Convergence • Auto Deploy • HW version 8 4 • Storage DRS • Profile-Driven Storage • VMFS 5 • Storage I/O Control (NFS) • Network I/O Control (per VM controls) • Distributed Switch (Netflow, SPAN, LLDP)

What’s New in v. Sphere 5. 1? • Single Sign On (v. CD, v. Shield, v. Center) • v. Sphere Web Client v. Center Server 5. 1 • Data Protection • Replication • v. Motion w/o shared storage • 0 Downtime upgrades of VMware Tools • Enhanced v. Center Orchestrator • v. Shield Endpoint • Storage DRS and Profile-Driven Storage integration with VCD • Enhanced Auto Deploy Availability Security Automation Compute Storage Network VMware v. Sphere 5. 1 • HW version 9 • 64 way SMP 1 TB VMs 5 • Storage Appliance • Storage Space Reclamation for VDI • Enhanced Distributed Switch • SR-IOV support

Compute, Storage, Network— Enhancements and Features 6

Auto Deploy Overview § Deploy and patch v. Sphere hosts in v. Center Server with Auto Deploy minutes using a new “on the fly” model § Coordination with v. Sphere Host Profiles Image Profiles Host Profiles § 2 new operating modes Benefits § Fast initial deployment and patching v. Sphere § Centralized host and image management § Reduce manual deployment and patch processes § Continue deployment even when a failure occurs 7

Distributed Switch Overview § Distributed Switch now delivers: § Network Healthcheck § Configuration Backup and Restore § Roll Back and Recovery § LACP Support v. Sphere Benefits § Visibility into physical and virtual network status § Backup and recover network settings § Fast recovery from lost connectivity or incorrect configurations 8

v. Sphere Scales to Support Mission-Critical Applications Overview § Create virtual machines with up to: § 64 v. CPU § 1 TB of v. RAM Benefits 2 x § 2 x size of previous v. Sphere versions § Run even the largest applications in v. Sphere, including very large databases § Virtualize even more applications than ever before (Tier 1 and 2) 9

Availability, Security, and Automation— Enhancements and Features 10

v. Motion (w/o Shared Storage) Overview § Live migration of a virtual machine without the need for shared storage § Extends VMware’s revolutionary technology for automated virtual machine movement Benefits § Zero downtime migration § No dependency on shared storage § Lower operating cost § Helps meet service level and performance SLAs 11

v. Sphere Data Protection Overview § New backup and recovery tool for the v. Sphere platform VDP § Replaces v. Sphere Data Recovery § Based on EMC Avamar VMware v. Sphere Benefits § Use less disk space with deduplication DATA DEDUPLICATED AND STORED ON VDP APPLIANCE *All editions and kits with the exception of Essentials 12 § Simple setup and management § Proven technology

v. Sphere Replication Overview Site A (Primary) § Virtual machine level replication by the v. Sphere host v. Sphere § Included with v. Sphere* v. Sphere Replication Benefits Site B (Recovery) § Low cost/efficient replication option § Simple setup from within v. Center Server v. Sphere § Integration with SRM enables automated DR process *All editions and kits with the exception of Essentials 13

v. Shield Endpoint Overview § Secure your VMs with offloaded antivirus and anti-malware (AV) solutions without the need of agents § Included with v. Sphere* Benefits § Simplified AV administration § Higher consolidation ratios by preventing the possibility of AV storms § Improved performance *All editions and kits with the exception of Essentials 14

v. Center Server— Enhancements and Features 15

Web Client Object Navigator Inventory Objects Create Custom Actions Overview § New, improved interface into v. Sphere delivers: § Browser-based experience § Custom tagging § Scalability § Enhanced workflow management Tabs Benefits § Platform independence § Tag based on specific business cases § Manage more objects and 3 x more active Portlets Add right-click extensions 16 sessions than ever before Sidebar Extension § Pause and resume even the most complex workflow or task

v. Sphere Web Client Interface Object Navigator Inventory Objects Tabs Create Custom Actions Sidebar Extension Portlets Add right-click extensions 17

Web Client—Native Plug-In Support 18

Single Sign-On Overview § Sign-on once rather than multiple times in v. Center Server Inventory Service v. CO v. Center Authoriza tion Active Directory Authentication (Single Sign On) v. Sphere Web Client v. Sphere Solutions v. Sphere Auditing Platform Services Open LDAP NIS Local OS Users Customer Identity Sources Benefits § Faster operations § Less complexity § Support for multiple identity services § Future building block for other VMware products and solutions 19

Single Sign-On v. Center Inventory Service v. CO Authorization v. Sphere Web Client v. Sphere Solutions v. Sphere Authentication (Single Sign On) Auditing Platform Services Active Directory 20 Open LDAP NIS Local OS Users Customer Identity Sources

v. Center Orchestrator (v. CO) Overview § Workflow Engine § Enhancements: § Web Client Integration (launch workflows) § New workflow design § Simplified configuration and installation Benefits § Execute workflows with a single interface § Simplicity thru drag and drop workflow creation § Automatic configuration § Deploy as a virtual appliance 21

Additional Features and Enhancements 22

The Best of the Rest § Platform • ESXi Platform Updates • New VM Features and Capabilities • Host Profiles § Network • Port Mirroring Enhancements • Scale § OS Support • Windows 8 Server and Desktop 23 § Storage • • • VMFS File Sharing Limits Space Efficient Sparse Virtual Disks 5 Node MSCS Cluster Storage Protocol Enhancements Storage Resource Management Enhancements • VMware v. Cloud® Director™ Interoperability **Details on the new v. Sphere Storage Appliance 1. 5 (which works in conjunction with v. Sphere 5. 1) are available in a separate customer overview

MEMORY © 2009 VMware Inc. All rights reserved

Memory – Host Memory Management Occurs when memory is under contention § Transparent Page Sharing § Ballooning § Compression § Swapping 25

Memory – Transparent Page Sharing 26

Memory – Ballooning 27

Memory – Compression 28

Memory – Swapping 29

Memory – Swapping 30

Memory – Ballooning vs. Swapping § Ballooning is better than swapping § Guest can surrender unused/free pages § Guest chooses what to swap, can avoid swapping “hot” pages § Idle memory tax uses ballooning 31

Memory – Rightsizing § Generally, it is better to OVER-commit than UNDER-commit § If the running VMs are consuming too much host/pool memory… • • 32 Some VMs may not get physical memory Ballooning or host swapping Higher disk IO All VMs slow down

Memory – Best Practices § Avoid high active host memory over-commitment • No host swapping occurs when total memory demand is less than the physical memory (Assuming no limits) § Right-size guest memory • Avoid guest OS swapping § Ensure there is enough v. RAM to cover demand peaks § Use a fully automated DRS cluster • Test that v. Motion works • Use Resource Pools with High/Normal/Low shares • Avoid using custom shares 33

CPU © 2009 VMware Inc. All rights reserved

CPU – Overview § Raw processing power of a given host or VM • Hosts provide CPU resources • VMs and Resource Pools consume CPU resources § CPU cores/threads need to be shared between VMs § Fair scheduling v. CPU time • Hardware interrupts for a VM • Parallel processing for SMP VMs • I/O 35

CPU – v. SMP § Relaxed Co-Scheduling: v. CPUs can run out-of-sync § Idle v. CPUs incur a scheduling penalty • configure only as many v. CPUs as needed • Impose unnecessary scheduling constraints § Use Uniprocessor VMs for single-threaded applications 36

CPU– Scheduling Over committing physical CPUs VMkernel CPU Scheduler 37

CPU– Scheduling Over committing physical CPUs X VMkernel CPU Scheduler 38 X

CPU– Scheduling Over committing physical CPUs XX XX VMkernel CPU Scheduler 39

CPU – Ready Time § The percentage of time that a v. CPU is ready to execute, but waiting for physical CPU time § Does not necessarily indicate a problem • Indicates possible CPU contention or limits 40

CPU – NUMA nodes § Non-Uniform Memory Access system architecture § Each node consists of CPU cores and memory § A CPU core in one NUMA node can access memory in another node, but at a small performance cost NUMA node 1 41 NUMA node 2

CPU – NUMA nodes § The VMkernel will try to keep a VM’s v. CPUs local to its memory • Internal NUMA migrations can occur to balance load § Manual CPU affinity can affect performance • v. CPUs inadvertently spread across NUMA nodes • Not possible with fully automated DRS § VMs with more v. CPUs than cores available in a single NUMA node may see decreased performance 42

CPU – Troubleshooting § v. CPU to p. CPU over allocation • Hyper. Threading does not double CPU capacity! § Limits or too many reservations • can create artificial limits. § Expecting the same consolidation ratios with different workloads • Virtualizing “easy” systems first, then expanding to heavier systems • Compare Apples to Apples • Frequency, turbo, cache sizes, cache sharing, core count, instruction set… 43

CPU – Best Practices § Right-size v. SMP VMs § Keep heavy-hitters separated • Fully automated DRS should do this for you • Use anti-affinity rules if necessary § Use a fully automated DRS cluster • Test that v. Motion works • Use Resource Pools with High/Normal/Low shares • Avoid using custom shares 44

NETWORK © 2009 VMware Inc. All rights reserved

Network – Load Balancing § Load balancing defines which uplink is used • • Route based on Port ID Route based on IP hash Route based on MAC hash Route based on NIC load § Probability of high-bandwidth VMs being on the same physical NIC § Traffic will stay on elected uplink until an event occurs • NIC link state change, adding/removing NIC from a team, beacon probe timeout… 46

Network – Troubleshooting § Check counters for NICs and VMs • Network load imbalance • 10 Gbps NICs can incur a significant CPU load when running at 100% § Ensure hardware supports TSO • Use latest drivers and firmware for your NIC on the host § For multi-tier VM applications, use DRS affinity rules to keep VMs on same host • Same v. Switch / VLAN, rules out physical network § If using Jumbo Frames, ensure it is enabled end-to-end 47

Network – Best Practices § Use the vmxnet 3 virtual adapter • Less CPU overhead • 10 Gbps connection to v. Switch § Use the latest driver/firmware for the NICs on the host § Use network shares • Requires Virtual Distributed Switch 4. 1 § Isolate v. Motion and i. SCSI traffic from regular VM traffic • Separate v. Switches with dedicated NIC(s) • Most applicable with Gigabit NICs 48

Key Takeaways – Performance Best Practices § Understand your environment • Hardware, storage, networking • VMs & applications § Advanced configuration values do not need to be tweaked or modified • In almost all situations § Use fully automated DRS § Use Paravirtual hardware 49

Tools – v. Center Operations Slide 50 § Aggregates thousands of metrics into Workload, Capacity, Health scores § Self-learns “normal” conditions using patented analytics § Smart alerts of impending performance and capacity degradation § Identifies potential performance problems before they start 50

Tools – v. Center Operations Slide 51 51