IBM Buys SoftLayer, But Will They Learn From Them?

IBM didn't just pick up a hosting company with their acquisition of SoftLayer this week, they picked up a sophisticated data center operations team -- one that could teach IBM Global Technical Services (GTS) a thing or two about efficiency when it comes to next-generation cloud data centers. Here's hoping IBM will listen.

Despite all the press and Twitter rhetoric, this is less a cloud buy than it is a technology and customer acquisition move by IBM. While on the surface, IBM SmartCloud Enterprise+ and SoftLayer may seem to be competing offerings, there is virtually no overlap here. IBM's public cloud/managed hosting offering is aimed mainly at IBM's existing large enterprise customers and designed more to accommodate their efforts to transition traditional applications to the cloud model. SoftLayer is a hybrid cloud service built mainly for web-centric small- to medium-sized businesses. Under the covers, IBM has been learning just how automated a cloud platform really needs to be, whereas SoftLayer started from the philosophy of "automate everything possible." Like most leading public cloud players, SoftLayer has programmed and scripted nearly all infrastructure administration functions via their unique cloud controller and made it accessible via API to their customers; they think manual provision is only a last resort. They also think commodity first, like most clouds do, but, like IBM, believe very much in strong base SLAs of availability and performance. This is why SoftLayer offers dedicated servers (bare metal and provisioned in hours) and a common high-speed dedicated network within and between their data centers. The way they approach and manage their data centers reflects this fast, cheap, and programmable approach. 

SoftLayer also isn't your typical hosting company, as SoftLayer blends hosting and cloud purposely. Their automation and management approach treats dedicated infrastructure and virtual, multitenant infrastructure as a common pool of resources.

OpenStack vs. CloudStack? No.

SoftLayer was one of Citrix' first and most vocal reference customers for both the XenServer hypervisor and their CloudStack IaaS software. IBM is strategically aligned with OpenStack. But this is not a CloudStack versus OpenStack story. SoftLayer's automation and management of their environment is not dependent on CloudStack. SoftLayer simply leverages CloudStack as one mechanism for VM deployment. SoftLayer already provides the OpenStack Swift object storage service as part of their offering (a legacy of the CloudStack deal). 

Where this could get messy is through the integration (a process a lot of companies who have been acquired by IBM call Bluewashing, which was also the basis for my creation of the term Cloudwashing). IBM said their planned integration model will leverage SoftLayer “as-is” via API integrations, so IBM can more quickly deliver SoftLayer services to market. Over time, however, SoftLayer will clearly become part of the SmartCloud family and will be included in a move towards a more unified architecture based around KVM and OpenStack. That likely will take the form of a new-generation cloud platform that launches side by side with the existing CloudStack-based service. This approach is not dissimilar to Rackspace's cloud strategy. The other Texas-based cloud leader launched their OpenStack-based cloud service next to their legacy Rackspace Cloud service. Customers can launch new workloads in either Rackspace cloud service but the clear go-forward platform is Rackspace Open Cloud. 

Which SmartCloud Cloud Is Right For You?

Given the different focuses, both in customer base and operational model between SoftLayer and IBM SmartCloud Services, it is likely the two will remain separate for some time. Likely, SoftLayer will be positioned as the preferred service for next-generation Systems of Engagement  — more of a true cloud  — letting IBM reposition SmartCloud Enterprise+ as more a hybrid architecture supporting Systems of Record. It would be wise to keep the focus of their integration efforts around the initial stated focus — between these services. As has been stated before in this blog, business services are not either System of Engagement or System of Record but are a hybrid of the two, and different parts of your business services will be better served by different deployment environments. Keeping this from being a problem requires strong integration between the deployment types, which customers will value. 

Integration between services, however, isn't a strength of either company. SoftLayer, as observed by Forrester Analyst Lauren Nelson in her Forrester Wave of Hosted Private Cloud Services, has two separate portals and a mix of API sets (which are mostly SOAP, not REST). IBM has long been accused of being well integrated only so long as you ask IBM Global Services to do the integration work for you. 

In moving both SmartCloud and SoftLayer to an OpenStack foundation, IBM has the opportunity to ensure strong integration while distinguishing each service and preserving the needs of their two separate and unique customer bases. IBM also has the opportunity to discount what made SoftLayer unique and treat it as simply a market consolidation move. This would benefit no one.

If IBM GTS plays their cards right, they will be a more relevant cloud provider through this move and one that appeals to a broader and more diverse customer base.

Save the Date - June 18, 1pm ET: Both IBM SmartCloud Enterprise and SoftLayer were evaluated in Forrester's Public Cloud Platform Wave, which will be published this month. On Tuesday the 18th of this month, Forrester VP and Principal Analyst John Rymer and myself will conduct a client webinar on this research and how each vendor fared in the comparison. You can register for this call here

- Lauren Nelson, Henry Baltazar, & Dave Bartoletti contributed to this blog post.

Comments

Riding the automation riptide

Nice post James, thanks. While the original service move by IBM was a genius strategy in early 1990s, the type of automation described here and in the cloud and web services generally seem to be an obvious, classic disruption for IBM given so many boots on the ground globally with such a strong misalignment with customer and market needs. It reminds me of a story this week on the head of Japan for IBM attempting to cut back on head count and now being sued for it, sort of summing up the challenge for IBM.

While there is still much to be done in IT, the new generation of technology is much smarter and more efficient, with more standards requiring less integration, and typically cost less as well. Indeed the definition of service model itself has changed from one that described human integration of incompatible IT systems to freedom from lock-in and misalignment in the entire IT cluster, internally and externally, to automation --in an industry that became addicted to incompatibility for its business model.

IBM is not alone, but I wonder how many boats will be tossed in the riptide between oceans, at what cost--how many casualties--particularly once global printing of capital is 'eased'. We can already see the trend in the quarterly reports of most of the incumbents--emerging markets are slowing while mature markets are set for low long-term growth, with many customers saddled with historic levels of long-term debt in an era that will likely (almost certainly) see higher interest rates.

I really wonder whether the entire industry is structured properly for the future--size, culture, product mix, skills, ecosystems, alliances... do not yet see much sign of acknowledging a fairly dramatic sea change underway with strongest riptide in a generation. Unlike previous generational change, these channels are now shallow, currents are swift, requiring a much different vessel, experience, and seamanship.