The conversation ought to be familiar by now.
Many IT organizations are taking a hard look at themselves, and are realizing they have to learn to compete for internal customers.
To do so, they need to start looking more like a modern IT factory, and less like your father's IT shop.
A successful transformation essentially boils down to organizing for success: reformulating many existing IT roles, and creating entirely new ones.
Transforming even a moderate-sized IT shop takes a unique brand of leadership. The journey isn't quick or easy, but it does offer substantial rewards to all involved.
And, to put it gently, in many cases your options to ignore or postpone the inevitable aren't all that appealing :)
New IT inevitably means new roles and new processes, which -- of course -- implies the need for newer tools that are oriented around creating easy-to-consume services vs. simply monitoring the underlying technical components. For example, most VCE Vblocks are shipped with UIM as an "orchestral enabler" that can expose infrastructure and allow it to be consumed as a service.
But not all clouds are Vblocks, and not all IT environments can be managed using a single paradigm.
Into this transforming IT world, EMC will be launching IT Operations Intelligence v9 early next week.
While the product has plenty to like on its own merits, I'd like you to consider this a preview of some of the bigger ideas finding their way into next-gen IT management tools -- both from EMC and others.
The New IaaS Operational Role
In the ITaaS world, it's all about service creation and service delivery.
Everything runs on infrastructure of some sort, so the person running the IaaS component (infrastructure as a service) is rather an important key actor in all of this. In today's world, the NOC operator comes the closest to this IaaS role in terms of roles and responsibilities.
Any infrastructure service is a composite abstractions of server resources (usually abstracted at multiple levels), network resources (similarly abstracted) and -- of course -- storage resources, also abstracted at multiple levels.
The people consuming the services prefer to think and visualize things in terms of "their" abstracted services, e.g. "my" applications, "my" infrastructure, "my" service delivery and resource usage, etc.
The IT professionals delivering these consumable, user-oriented services have to be able to map those composite abstractions down into the entities that are providing them for any number of reasons: understanding configurations, performance problems, resource utilization, etc.
Perhaps the best way of thinking of the EMC IT Operations Intelligence product is helping the IaaS ops person do their (new) job by being able to jump back and forth between multiple abstractions, and correlate them as needed to solve a particular challenge.
Hey, if it was me, I'd call it "infrastructure service delivery manager" or or maybe "abstraction manager", but I didn't get to vote on this one ...
Now that you understand the intended use case, let's look at what's new and innovative in this product. There's a lot to like.
Abstraction Management
If you've ever worked in the network domain, you're comfortable with multiple layers of abstraction.
End user sees what looks like a point-to-point connection; reality is layers upon layers of managed network services. Some of these are real networks, more recently these segments now include things like VMware's distributed switch, or perhaps a Nexus 1000v.
The same is largely true in the application and server world: an "application" is typically multiple related entities, all with virtualized server resources, all which eventually land on a physical bit of server at some point. Ditto with storage: apps see logical storage which, in turn, is usually abstracted multiple times before we get to a physical device of some sort.
What's new about the IaaS role is that the service delivery manager has to correlate *all* of these abstractions simultaneously. Logical app group is formed from logical apps formed from individual tasks, each with their own complement of abstracted server, network and storage.
So, at a minimum, you need a couple of things.
- The ability to define abstractions that are visible and meaningful to the consumer, e.g. "my stuff"
- The ability to map down the components through multiple layers until you get to some sort of physical device
- Use that abstraction and correlation to do useful work across many domains: solve a problem, ensure that things like VMotion can work, run a report, etc.
So far, so good.
Management Mashups Go Mainstream
Much as we might often try as vendors, the world's IT management challenges can't be effectively solved with static, pre-configured functionality. Any serious IT management approach requires the capabilty to extend functionality in a number of directions without the direct involvement of the vendor: adding new types of devices, creating new management applications, and so on.
Historically, a lot of that has been done with a combination of scripting and lower-level coding.
If you've spent any time with the progressive #devops community -- whose mantra is the ever-appealing "infrastructure is code" -- they prefer to work with exposed REST APIs, and a variety of tools for rapdily composing functionality and creating customized management apps.
There's a much broader discussion to come soon, but you get your first "sneak peek" at EMC's "mashable infrastructure management tools" with this product.
Please stay tuned for much, much more :)
Delivering Services Means Delivering Service Portals
We have this strong belief that -- in the future -- most IT will be delivered as a service -- whether internally, externally or more likely some combination. That means exposing capabilities to others that are easy to consume, easy to understand, easy to manage and so forth.
In essence it quickly becomes important to get good at exposing portals. Potentially lots of them, each aligned around what your individual "service customer" wants to see. In this world, there is no "best portal", except in the eyes of the person using it.
Although the Infrastructure Operations has some decent out-of-the-box portalling capabilities, where it gets *very cool indeed* is when the basic capabilities are considerably extended with the mashup capabilities.
Developing an entirely new view of relevant infrastructure information (and making it easy to consume!) becomes something you do in a few hours vs. a few weeks or more.
From my point of view, this is something to pay attention to.
If you're an IT service provider, your ability to whip up any custom view your customer might want to see becomes an enormous competitive differentiator especially when selling to enterprise IT groups who tend to have "their way" of doing things. You can do your own portal mini-apps, offer to do them for a price, or -- potentially -- put the mgmt app development tools right in the customers' hands over time.
If you're an enterprise IT organization transforming to look like a competitive internal service provider, this will be a big deal as well. Why? Everyone wants "their view" of the data, and IT's ability to gen up consumable and customized views of configurations, performance, costing, availability, etc.
When selling any form of IT services (in the SP sense), everyone wants to know what it will look like to them. Indeed, in working with customers, seeing is believing when our EMC folks whip up a cool management portal applet in nothing flat :)
There's More, But ...
I really didn't want to go into a blow-by-blow description of the new product -- more of that detail is coming when the product is officially launched next week. There's a heckuva lot to consider, and it takes a fair amount of time to process and understand.
What stood out to me -- above all -- is that this new product is an excellent example of what's going to be needed to support these new, converged roles and workflows that are starting to rapdily show up in ITaaS models.
A final thought: if you're looking at this and thinking "gee, this isn't how we do things today!", I guess that's the point :)
The business model the publishing industry (et. al.) have followed is pretty much along the line of "The Cabots speak only to the Lowells, and the Lowells speak only to God." The customer's main function is to STFU and take what's shoveled out. Feedback is neither desirable nor encouraged. And actually ASKING the customer what he/she wants? Actually TELLING them directly what's coming out by asking them
Posted by: Doug Bell | January 21, 2013 at 12:53 AM