Early Cloud Investment makes for a Buyers Market

The $250mil HP-Microsoft deal announced a couple weeks ago speaks to how quickly and seriously the big players have converged on the Cloud opportunity.  Literally before the market has even come to understand what 'Cloud' means, we've seen Cloud deals involving many of the top global IT players including Cisco, EMC, VMware, HP and Microsoft.  Only Oracle-Sun is dialing back their cloud investment.

This is different than the uptake around ASPs many years ago.  Small pure-play start-ups like Corio and Jamcracker were prominent in the market for years with little competition from top IT firms, who instead chose to sit out the market wave and acquire ASP capabilities after the bubble burst. This early involvement of larger players means more vendors with capabilities and expected staying power; in effect two Savvis' for every one GoGrid.  So customers don't have to choose between start-ups who "get it" and dinosaurs who don't.

With more investment, sooner, we can expect the life cycle of the category to be similarly accelerated.  It bodes well for customers looking for validation and future-proofing from leading vendors.  It will probably also shake out the smaller players sooner, as they find themselves competing with established tech brands touting seemingly less-risky solutions.  Perceived security and 'trust' will be key aspects of success, especially in the hosted cloud space.

So with more M&A and investment expected relating to the Cloud, take your time before making big investments, plan with an eye towards federating your external Cloud services with your internal data center environment (rather than creating a hodge-podge of cloud silos), and enjoy a developing buyers market in all things Cloud.

Best Practices for Choosing an External Cloud Vendor

The idea of Cloud is catching fire.  One vendor recently shared that half of their customers are considering external Cloud storage providers. (As this was from a backup software vendor, I'm not sure this is such good news for them). 

The good news for you is that every IT vendor out there is delivering products or services to meet this growing Cloud interest. And the bad news is... every IT vendor out there is delivering products or services to meet this growing Cloud interest. Many are the same products that were last year's "Green IT" solutions. Some are enterprise-class utility computing products aimed at your data center. Some are public web-based services for SMBs.

I've been working now for some time on new External Cloud product efforts, and am seeing the potential confusion just in this one area of Cloud computing. So I thought I'd share a few tips I've picked up that may help you as you are considering new projects in "the Cloud".

1. Like any outsourcing decision, start with projects and applications that are not mission-critical or performance-sensitive.
2. Spend time to develop an architecture to address the linear scalability, parallel processing and distributed data aspects of cloud computing.
3. Understand how your management, monitoring and policies will accommodate external storage and compute resources especially in terms of security, availability, utilization and compliance.
4. Compare Cloud subscription costs with your burdened total costs of ownership -- Cloud should deliver savings for short-time horizon projects, where specialized admin staffing would otherwise be needed, and where large fixed investments may not be fully utilized.
5. Stick to annual contracts especially as Infrastructure as a service will become commoditized over time.
6. Expect that only a few leading ecosystems will emerge once the hype passes – choose your vendors carefully looking for a track record, customer references, a documented SLA and top quality infrastructure.

Good luck with your projects, and feel free to share what you're learning in the Cloud.

Storage in 2010 - Seeking Efficiency in the Cloud

I came across a useful story that compiles a number of the 2010 outlooks for the storage industry.   Forrester, ESG and Symantec seem to be sharing an optimistic view, with spending to increase while customers seek ways to be more efficient.  This points to increased consideration and adoption of hosted storage or Storage-as-a-Service offerings this year.

Storage-as-Service has been an established, viable offering well over the past decade.   Companies have been using hosted storage services or ‘Cloud Storage’ going back to at least 1993 from service providers like LiveVault (now part of IronMountain).  Like any outsourcing activity, IT organizations leverage out-of-house hosting providers to lower their capital expense as well as administrative staffing requirements. With the increased adoption of larger bandwidth connections across corporate WANs, and the continued growth in data, we can expect to see more offsite storage useage in 2010 and beyond.

Storage-as-a-Service offerings will grow in popularity from off-line archiving to online product use over the next 5 years.  Currently Cloud Storage is mostly limited to offline uses where the network latency isn’t an issue, such as for video and email archiving, and we will predominantly see this type of use for Cloud storage in 2010.  As network connections between enterprises and their service providers are improved along at least a few key dimensions, we will see these use cases expand to include more production-oriented data storage.  Over time Cloud storage will become an alternative to nearline storage, and some service providers have been quick to launch services to begin this evolution.

However, at least three major advances in networking will be required to enable the broader growth of the Storage-as-a-Service market.   Moving data over a WAN link is a major inhibitor to production storage service use.  Transit acceleration will need to be a core part of the solution to improve the performance to be closer to native LAN latencies.  Similarly, the availability of the connection needs to be better maintained, and more advanced techniques will need to be employed to assure more highly available connections between the datacenter and the service provider, which will also improve throughput. Finally, security for both data-in-transit and data-at-rest will need to be in place to assure that business and customer data cannot be breached, especially within multi-tenant hosted environments.

Cloud Stepping into Commodity Pricing Trap?

Back before the holidays there was a story about Amazon bringing out demand-based pricing for their EC2 service. This 'spot instance' program, still in beta, has already been successful in building some awareness and blogosphere traffic for them.

When I initially read the story I wrote this off as a bad trip down "Revenue Optimization" lane. The self-destructive road that the travel industry took in the 90s. Being part of that business, it was exciting to develop systems to determine optimal pricing that mapped to demand segmentation. It became a real craze across airlines, then hotels. And it's still a big feature of that business today. But as the economy soured, the use of technology to maximize revenue became an accelerated exercise in optimally dumping excess inventory. What we later learned was that when travelers compared notes and realized that one paid a hundred dollars less than the other for the same seat or bed, it had an incredible commoditizing effect on how they viewed your product. And fueled by the power of the Internet, it's a trap they never escaped. We perpetuate this legacy every time we begin our travel planning by using a price comparison site like kayak, orbits or travelocity. The brand value of the individual airline, hotel or rental car has all but vanished. The same attraction recently hit Broadway – the phenomenon was captured nicely in a NYT article.

But then I went back and read the details of how the Spot Instance program works. Turns out you bid a rate, and if the price falls to that level, then your Amazon instance will turn on. And just as quickly, it will turn off as soon as the spot price exceeds your maximum price! It would be like kicking the discount vacationer out of their bed because a full-rate business traveler showed up at the front desk. And who has a workload like that, anyway???

Time will tell how this experiment plays out, but I’d suggest keeping an eye on this service, as we are either seeing the next chapter being written in the Pricing and Revenue Optimization text book, or the emergent Cloud industry starting down the commodity road to ruin.