Summary at the top = nice View 5 feature overview and then HUGE stuff around using vCOPs to monitor View environments. This is critical stuff for building out and maintaining large VDI deployments.
- Found that a lot of the underlying requirements for Server Virtualization are the same as for VDI (stability, speed, scalability, etc.).
- First VDI solution to enable full 3D in Windows without any hardware….can do it all in software.
- Easily enable 3D apps for users
- Storage Access Optimization with Content Based Read Cache
- In-memory cache of common block reads
- Applicable to all types of desktops
- Completely transparent to the guest
- 100% server based
- No special array technology needed.
- Lower storage costs by reducing peak IOps
- This is huge for boot storms….will cache I/Os at the ESX level.
- Ensure consistent user experience by handling worst case peak usage.
- vShield Endpoint
- More benefit to storage I/O
- Publicly announced commitments from….
- Trend
- Symantec
- Kaspersky
- Sophos
- Glad to hear commitments…it’s taken quite long enough for anyone but Trend to support it.
- Higher security as AV module is isolated and more difficult to evade.
- vShield App in the future for better physical network protection.
- Context aware in enforcing security with vMotion, etc.
- Transparent, application aware firewall.
- Talked about this morning in the General Session.
- vCenter Operations for View
- End-to-end monitoring for vCloud environments using vCenter Ops
- Now a live demo about this….
- Looking at the entire solution..who is involved?
- VI Admins – build and configure ESX, on/off-board users, Maintain healthy infrastructure, guarantee good end-user experience.
- Desktop Specialists – provide 1st level technical support, perform basic triage, engage VI admins when needed.
- Helpdesk – log and manage tickets, hotline for end-users.
- BUT….end-users just want to do their work!
- Typical problems in the end-to-end View infrastructure
- Client side — using iPad from home….network issues.
- Internal network
- ESX configuration.
- View-level config issues (not enough VMs provisioned, etc.).
- Storage — lack of storage and/or backend bandwidth.
- vCenter Operations for View — “Divide, Correlate, and Conquer”.
- Gather thousands of metrics to measure performance in real time.
- PCoIP stats (user experience), IOPs, r/w latency, etc.
- Breaking up into sections watched by vCOPs for…
- Clients
- Network
- vSphere
- Storage
- View Application
- How does it work?
- Learns the dynamic ranges of each and every performance metric.
- Learns behavior patterns and identifies abnormalities.
- Correlates multiple anomalies into a Health Score
- Isolates root-cause for performance degradations.
- Automatically – no configuring/tuning needed.
- Walkthrough of an end-user call.
- Can search by user.
- See lots of info about the user/VDI session/etc.
- Red/yellow/green scoring to highlight any issues.
- Canned video walkthrough
- Discussion about how vCOPs can and does alert before end-user visible issue happens (based on robust trending/alerting).
- Even further futures – enhanced vSphere web client that is really a mashup with vCOPs, has specific stuff even for View admins too.