The term IT-as-a-Service has been coming up a lot lately, in a blog article Read Write Cloud by Audrey Watters , the one she refers to at Dev Central by Lori MacVittie, and in keynotes from 451 Group analysts at the recent HCTS conference I attended.

I think they've got it right - which is why we have tried to provide these services from ENKI since it was conceived.  However, just like with the definition of Cloud, there is a lot of confusion around what IT-as-a-Service actually is. 

For example, Audrey has a block diagram showing monitoring and metering added to cloud in front of a push-button interface, but this isn't IT-as-a-Service: it's simply an improved cloud with some of the things that are currently missing from it added in - essentially the infrastructure 2.0 that Lori is speaking about.

If you have worked in a reasonably-sized company and used their IT department to support your work, what strikes you is not that it's NOT pushbutton but that is IS full service.  You do your job - which is generating value for the company - and they do theirs - which is deploying and managing the apps you need to do your job.   That's IT - as a service.

The advent of cloud technology has made this function in a company look dated because in the past it was slow to provide results, or expensive.  But the needs it serves haven't changed.  For example, the idea that the marketing department would manage its own servers in the cloud would seem ridiculous to them, even if some smaller companies do this now to save money.

I think the confusion about this is illustrated by one of the commenters on Audrey's blog, who wonders why IT-oriented folk would want IT-as-a-Service.  However, the customers for this service aren't IT folk; they're business people.   That is the ultimate direction of Cloud Computing: to democratize access to IT services.   And today's focus on the self-service cloud has gone part-ways down that path, but it can't get much further if it remains self-service, because it still requires someone to serve themselves from a resource that requires architectural, systems integration, systems analysis, and systems administration skills to use successfully: in other words, an IT department, or in smaller companies, a few "IT guys."

Where I think Cloud is going is that it will become a supporting technology for a revolution that will eliminate the need for an IT department in small to medium enterprises through on-demand services, and provide a powerful technology platform to make an IT department more effective in larger enterprises.  Today's cloud will stop being a product and become an enabling technology in a stack that provides IT-as-a-Service, or as Lori puts it, a "parfait."

Here at ENKI we use NetSuite for our back end accouting and CRM systems.   If you read our bios, you'll see that many of us are NetSuite alumni, and we like the "one system" approach that it provides for us to manage the business and in particular, to guarantee customer support responsiveness.

But, is NetSuite cloud computing?   Today, I logged into NetSuite and saw a little banner that said, "See Why CNBC Says that NetSuite is at the Center of Cloud Computing!"    As a 4-year veteran of providing cloud computing to customers, an attendee at numerous panels and events, and the writer of blogs and papers on Cloud computing, my first reaction is, "NetSuite is not cloud computing!"   The problem is that now 4 years into the Cloud Revolution, we're still debating what Cloud computing really is.   The consensus of the pundits - renewed at last weeks's 451 Group HCTS conference by Rachel Chalmers - seems to be that it has 4 fundamental characteristics:

1. Computing delivered over the internet.
2. Scalable, i.e. you can add or reduce resources to meet need - in particular to assure a constant level of service.
3. Pay-as-you-go, meaning that you only pay for what you use.
4. Self-service, meaning you can add or subtract resources/services yourself - though this is not universally agreed-upon.

I just spent the last three days in Las Vegas with Tier1 and 451 Group Analysts and their hosting, datacenter, and cloud provider customers.   This was my first time as ENKI was invited by one of our customers, CloudSoft, who was a sponsor of the event.    HCTS is an opportunity for Tier1 and 451 Group customers to market to other customers of the analysts, and of course for the analysts to share their insights and market their services to guests such as us... but most of all, it is an opportunity for people to network!    I thought I's share a little of what I learned.

First of all, the technology to provide internal and public clouds is mature enough for relatively easy (compared to a few years ago) deployment of clouds, whether by companies for their own users, or by managed services providers.   With the availability of the new VMWare vCloud Director (which ENKI is about to release) as well as other products from cloud.com, OnApp, or 3Tera (now CA) - all of whom were exhibiting at HCTS - I expect that the number of clouds deployed around the world will grow rapidly.   In particular, 3Tera's team seems energized by the CA acquisition, and poised to deliver a spate of new and solid enterprise features.

On the other hand, there are significant remaining technical challenges to making those clouds available to users with a wide variety of use models, so I also expect that many of the companies that try to field clouds, even with the new tools, will struggle to make them cost-effective or deliver attractive SLAs.   And as the customer base for cloud increases in size, IT departments and managed service providers will face significant challenges which the current crop of cloud deployment systems don't address.  In particular, reducing system administration effort for IaaS, and working around storage and interconnect technology that still hasn't caught up to cloud workloads are all challenges that casual cloud shops won't be able to overcome.   So, for the time being, there seems to be little substitute for experience and solid IT background if you're deploying a cloud - but at least you no longer have to code it up yourself!  One of the technology surprises in this area was Microsoft, which plans to offer Azure as a combination hardware/software package for companies to deploy under their own roof.

Because of these technological changes, as well as sufficient experience under the belt of the cloud computing industry and customers in attacking problems of security and certification, the analysts reported that enterprise adoption of public cloud services seems well underway.   There are still plenty of enterprises that are shying away from cloud for these reasons, but a significant fraction of enterprises are now deploying cloud solutions for core application needs, not just development and test. 

One other observation that I had is that 451 and Tier1 have a lot of investor clients with money to invest in this industry.  Both organizations have analysts whose job is to supply the financial community with intelligence on where to invest their money.  The result was that we were surprised with a deluge of interest by investors in ENKI.    This influx of money is going to continue to develop new cloud solutions as well as fueling diversity in cloud product offerings that will be proving the dire predictions that there will only be one or two giant cloud computing vendors within a few years.   As one of the analysts pointed out in session, there is enough opportunity for cloud vendors that the big companies can't soak it all up, and can't address the entire spectrum of customer needs.   As our CEO Dave Durkee has written about (please request his White Paper from us!) there will be a lot of cloud solutions other than the "race to the bottom" price-focused offerings that the big names are coming up with.

This morning I read a blog by Ann All over at IT Business Edge asserting that (enterprise) cloud users who consider cloud computing as outsourced IT was "troublesome."   It's an interesting blog (as well as the one that is referenced by the highlighed "troublesome") and I recommend it.   However, I don't agree with it, even if I put myself into corporate IT shoes.

I think the first wave of cloud adopters were the "garage entrepreneurs" who didn't have an IT department.  The next wave - which we're in now - is the intrapreneurs who have no use for their IT department and were alienated from its slow and excessively centralized approach long ago.   This feels right to me since I was one of those intrapreneurs back when I worked at HP and couldn't get our internal IT to help with a rush CRM project.   Not only did I have to put together my own data center, but I also had to hire and manage a 24x7 operations staff to keep the application up and running.  I was wishing I didn't have to, the whole time!

So, I disagree that it's "troublesome" that today's corporate cloud users see Cloud Computing as IT Outsourcing.  They see it that way because that is what they need as intrapreneurs includes more than just server boxes, encompassing the entire range of services an IT department would offer - but they need it on-demand.  This may trouble corporate IT, but if my experience is any guide, they were already troubled beforehand, since HP's IT department fought my little island of IT when I built it - even though they couldn't offer what I needed.

In fact, I think IT should be heartened by the fact that intrapreneurs and a few Cloud providers see Cloud as outsourced IT.  The reason is that the next wave of Cloud users will be mainstream business users who need temporary or rapid deployment resources, and IT will have difficulty offering their services to these users cost-effectively.   Instead, IT will have Cloud vendors already trained and prepared to partner with them as an outsourcing adjunct to their services.   Judging by what I see in many larger corporations today, IT is already very familiary with outsourcing and knows how to manage it well.

In the meantime, SMEs will benefit from Cloud as outsourced IT since that will provide them far better service than they were ever able to get either by going the DIY route or traditional IT outsourcing.

I just enjoyed a few days at VMWorld, walking the booths and meeting other vendors.    I go to see what technologies should be incorporated into ENKI's services but also to learn where the trends in the industry are moving.  As it often turns out, those are the same things, since ENKI is always striving to provide the best in cloud computing.

Of course, the theme of the conference is Cloud Computing.  And the big star of the show - not a surprise considering who sponsors it - was VMWare's new vCloud Director product, known to many as "Redwood."   Director for the first time enables end customers of cloud providers to have self-serve access to the majority of VSphere's features, as opposed to the far less capable VCloud Express product which anchored VMWare's cloud infrastructure offering so far.   We've been playing with Redwood in beta, and are impressed.  ENKI plans to release it to our customers as soon as possible after its general availability in October.   Director will allow our customers to combine self-service for their casual applications (test, staging, development) with our traditional fully managed services which are aimed at reaching the highest reliabiiity levels.

However, what interested me even more at the show was new technology that was being delivered to solve two of the cloud's persistent problems: high performance storage and security.   

Storage has always been a challenge for us, since the wide variety of loads and huge working sets presented by cloud customers in a computing cluster overwhelms traditional storage solutions.   Vendors have responded by simply beefing up their solutions, at best a stopgap.  ENKI currently uses Sun's Open Storage 7000 systems because we felt they are the best of the stopgap solutions.   However at this show a number of vendors are solving the problem in a cloud-centric fashion.  I've always maintained that storage caching should happen at the point of use, not in the storage system.  One vendor, Amplidata, has come close to the perfect architecture by relying on local storage - including SSD - in each compute node to do the caching, as well as scaling the storage by adding new independent units in a network infrastructure to meet increasing demand.  This effectively removes the central bottleneck in storage that Dave wrote about in his recent ACM paper, "Why Cloud Computing Will Never Be Free" (available on request.)   We're going to be taking a much closer look at Amplidata.  

Security has been a continuous concern for our potential customers saddled with compliance requirements.  ENKI's position has always been to provide a cloud datacenter that is built as much like a corporate data center as possible to allow the same type of security measures to be used in our cloud as our clients would normally use in their own datacenters.  However, the multitenant nature of the cloud has always raised the specter that despite such design, data sharing and cross-customer security issues will get in the way of HIPAA, SOX, or PCI compliance.   One vendor I saw at VMWorld, Catbird Software , offers a suite of products that will scan your cloud implementation once or repeatedly to verify compliance with regulations.   It can also be used to compare compliance in your traditional datacenter with that in your cloud implementation.  Obviously, it isn't going to validate the people-processes that make up many of the compliance concerns, but by validating the configuration and setup of customers' virtual private data centers, it goes a long way towards making cloud a viable alternative to the long, slow, expensive slog through implementation with your own staff and infrastructure.  We're looking forward to deploying it for our customers concerned about compliance.

In the next weeks I'll talk a bit more about some of the other products - especially storage - which I saw at VMWorld