Gartner – Emerging Technologies

June 24th, 2009 by Julia Lim

Greetings on Day 2 of the Gartner Infrastructure and Operations Management Summit in balmy Florida!

A day in and it’s pretty obvious there are some mega-trends running through the sessions.  Cloud computing is being covered extensively – and the underlying issues of reducing opex (power and cooling costs in particular) and getting away from traditional “overprovisioning” of servers crops up again and again.

But the star so far must be Virtualization. We’ve heard: Virtualization leads to cloud computing. Virtualization drives/enables/requires [Fill in the blank]. Moving beyond the initial server consolidation and test environment usages, virtualization is a driving/enabling technology that will have far-ranging impact on IT as evidenced by the emerging and disruptive technologies talked about in this morning’s keynote.

Analysts: David Cappuccio, Carl Claunch, Bob Hafner, David Williams

Carl Claunch:
1) Virtualization is not a new idea but a hot topic because it’s spreading into so many different areas/possibilities. Examples:

• aggregation – using virtualization in an way opposite to what it’s currently well known for; instead of breaking up a server into multiple machines, put together multiple machines and make them look and act like one
• new software distribution method – cutting down labor during installation and managing better version/platform challenges

2) Converging Communications – for interconnect, storage, data, video, voice, sensors, etc
- Everything IP
- Analog to Digital
- Integration

Yes it will be more complex to manage – the point is to “add complexity internally so that it reduces complexity where it matters more.” [Note: pretty much sums up our own product philosophy]

Bob Hafner: Previously we thought of the network as a pipe, with components built sequentially and often underutilized. Going forward…
1) Virtualize the network – do this logically instead of physically.

Benefits:

• reduced the number of boxes
• significant reduction in power, cooling and space
• improved performance – reduced congestion and latency

Issues:

• heavy use of VLANs
• requires better monitoring tools that can handle this
• requires good operational discipline

2) Virtualize UC and Communications-Enabled Business Processes

By 2013 – 40% of knowledge workers worldwide will have abandoned or removed their desk phones.

Video can be an issue – you don’t want the scenario of everyone with Webcams on their desktops. If successful, the network cannot yet handle this kind of widespread adoption not to mention the cost. Also, video has to be given priority over data on the networks – otherwise degradation of service. AND all this stuff needs management. [and we plan to have an announcement about this soon]

David Williams: Economy is another driver to adopt virtualization – cost savings coming from hardware, power and cooling, but management is always behind the curve.

1) IT Process Automation/Run Book Automation
4 companies 5 years ago – over 22 now
If you’re not ready for this, you should get ready because you will have this. All the big vendors [and ScienceLogic!] are embedding them. For the frameworks – RBA was originally a way to orchestrate all those tools they offer in their portfolio (all those acquisitions and not a lot of integration). Most IT orgs not ready for this; RBA works best across silos but most companies still work in silos. Success will depend on designated process owners – but that is currently very rare.

Ideal scenario example:
Standardize configuration. Automation process created to do configuration checks. Voila efficiency.

2) Behavior Learning Tools

• self-learning; ability to identify patterns
• dynamically changing thresholds – set against “normal” operations
• predictive capabilities gained by correlating and analyzing real-time data against historical and thresholds
• ability to apply learning at both element and services levels

Skepticism is the biggest competitor here. Virtualization is pushing the need for this – increases the amount of data/speed of data and need to do data analysis to help get to root-cause faster.

The RBA Path:

Discovery (also via integration) >> real-time data collection >> data modeling/profiling >> pattern matching and correlation >> dashboards and reporting

3) Performance Management Database (PMDB)
IT Infrastructure Usage Optimization

“If you thought the CMDB was hard, that was a dream compared to this.”
No one has these yet. CMDB – snapshot once a day. PMDB – snapshot once a second. Virtualization requires this frequency and granularity.

• aggregating, normalizing and anlyzing performance usage information from multiple sources (e.g., servers, netwrks, databases)
• delivering performance usage information to allow IT elements to be configured and provisioned in support of IT services

Popularity: 1% [?]

Related Posts

• No Related Post

Add comment