Trident Capital Blog

Thoughts on IBM’s Smart Analytics Cloud Announcement

Yesterday IBM announced the Smart Analytics Cloud, a cloud-based BI solution.  This solution will initially be available to IBM’s employees who, according to the announcement, will be able to access 1 PB of structured and unstructured data stored in 100 different data warehouses and marts.  Eventually, the solution will also be made available to external clients so that they can use it to develop their own cloud-based analytic applications.

While using SaaS BI to further its growing analytics business, IBM’s work on the Smart Analytics Cloud has the potential of helping the field in a variety of ways.

1. Market validation.  For those of us involved with the financing and development of SaaS BI this is a very important announcement because it validates the market and its potential size since IBM only enters what it feels are multi-billion dollar markets.  In fact, IBM may have realized that SaaS BI is a “killer app.”  I expect that IBM’s entry will help companies that were “on the fence” regarding the use of SaaS BI solutions to seriously consider them, thus helping all vendors working in this area.
2. Security and compliance.  CIOs continue to view security and compliance as two of the major inhibitors to the enterprise’s broader use of cloud computing in general, and SaaS BI in particular.  IBM has the opportunity to address these issues and make private clouds more secure.  More importantly, its involvement in the area will make CIOs feel more comfortable about using the cloud for mission-critical applications, BI being one of them.
3. Scalability.  For starters, IBM will allow 200,000 of its employees to access 1 PB of data through the Smart Analytics Cloud.  This will make it the biggest cloud-based BI solution in terms of users.  Understanding usage patterns, data movement from on-premise databases and data warehouses to the cloud, and query types and their impact on the Analytics Cloud’s performance can provide invaluable lessons to the developers of SaaS BI solutions around issues such as multi-tenancy, I/O, etc.
4. Data management.  The Smart Analytics Cloud will be built around Cognos’ software and mainframe hardware.  It is not clear from the announcement whether IBM will also manage the data in the cloud, like the pureplay SaaS BI companies do, or whether it will assume that all the accessed data warehouses remain on-premise.  I hope it’s the former since managing 1 PB of data stored in 100 different cloud-based data warehouses will offer the opportunity for some interesting lessons regarding how big data moves to and from the cloud, how this data is ingested and organized and becomes ready for analysis across 100 different data stores, etc.   Of course, while we know that 1 PB will be accessible for analysis, we don’t know how quickly the underlying data will be changing.  Will it be changing daily, weekly, monthly?  Are the changes multi-terabyte data sets themselves or something smaller?  IBM hasn’t said anything about developing a cloud-resident version of DB2, similar to Microsoft’s cloud-based SQLServer.
5. Architecture innovation.  Though essentially a private cloud, the scale of the Smart Analytics Cloud in terms of the number of users it serves and the size of data it contains will offer the opportunity to develop novel system and data architectures.  These architectures can then form the basis for the version of the Analytics Cloud that will be offered to external customers.  I’m hoping that IBM will offer the external version of the Analytics Cloud using a multi-tenant architecture as it can also innovate in that area both in terms of the technology but also in terms of the business model.
6. Technology innovation.  In the same way that Yahoo’s work on analytics for online advertising resulted in the development of Hadoop and other technologies that are now being used by a new generation of BI companies, through its work on the Smart Analytics Cloud IBM has the opportunity to develop innovative technologies for SaaS BI in areas such as data management, security, parallel query processing, data movement, data center architecture and management, etc.

By announcing the Smart Analytics Cloud IBM validates once and for all the SaaS BI market, has the opportunity to address some of the inhibitors to the broad adoption of cloud computing, and can create significant technology innovation in several areas.  It won’t be a quick win since cloud computing is still a very new platform.  But it can be a big win for IBM and all the other companies working in SaaS BI.

Notes from the Goldman Sachs Cloud Computing Conference

Last Wednesday Goldman Sachs held its annual software retreat in NY.  The topic this year was cloud computing.  This year’s program consisted of a series of panel discussions around the hardware, communications, hosting, services, and software aspects of cloud computing and SaaS.  Noteworthy:

• Each participating panelist provided their own definition of cloud computing that was aimed at showcasing each vendor’s cloud computing solution.  My sense is that today cloud computing is a “vendor push” solution rather than a “market pull” solution.  As a further validation of this statement I offer the opinion of a few CIOs participating in the event who stated that even by the end of 2011-2012 at most 15% of IT loads will be executed in the cloud, including SaaS applications.  This number is consistent with the one recently published by Goldman Sachs after they surveyed several CIOs on this issue.
• The early adopters of cloud computing are the companies that have already determined that IT is not one of their core competencies, and have the least IT to amortize.
• CIOs will use the cloud to:

1. Better control the expenses of running core applications, i.e., they won’t have to invest on IT infrastructure to be used only during peak conditions, .e.g., Black Monday for e-commerce retailers,
2. Save on resources that are used to run non-core applications,
3. Run SaaS applications, and
4. Achieve higher quality of service for business unit users at lower cost.  Through cloud computing CIOs would like to provide business units with the opportunity to decide how much to spend on running a particular application and under what SLAs.

• Two years ago the discussion was whether the enterprise will adopt any SaaS applications.  Today the conversation is about which on-premise applications will be replaced by their SaaS equivalent.  SaaS has penetrated the enterprise, is showing strong ROI which is in turn resulting in accelerating adoption.
• Infrastructure as a Service (IaaS) requires management skills that most IT shops today don’t have.  Most IaaS efforts are still in testing environments rather than production environments.
• Security, compliance (e.g., PCI, HIPAA), data transmission and data center connectivity are some of the biggest issues that will continue inhibiting the adoption of cloud computing in the enterprise.
• CIOs don’t want a single, ubiquitous stack for cloud computing but the ability to work simultaneously with multiple stacks coming from different vendors.
• The capital requirements of cloud computing startups will be high as they will look to create offerings and value propositions that can effectively compete with those from large IT vendors, e.g., IBM, Cisco, Microsoft, EMC, and be adopted by CIOs.