Recently Splunk announced the availability of its cloud offering, which is just further validation that large enterprises, along with the rest of the world, are moving from on-premise to cloud solutions. In this post I’ll share a few thoughts on why Sumo Logic started in the cloud and raise a few questions about Splunk's announcement as it relates to the broader machine data search and analytics market.
1. Why Sumo Logic started in the Cloud? I joined Sumo Logic with first-hand experience trying to shift an on-premise software company to a cloud agenda. The reality is that if you look at the traction and growth in every major category you will see cloud vendors rapidly taking share from the much larger on-premise incumbents. Salesforce.com started it, but recent examples include ServiceNow which just went public, SuccessFactors which was bought by SAP, and Workday, which is in hyper-growth as evidenced by its recent S-1 filing. What makes this transition so hard?
- On-premise vs. cloud architecture. Most on-premise software companies make the classic mistake of trying to port their legacy architecture to the cloud so they can take advantage of work and features that have already been done. In the majority of cases, this just does not work. Even the newest on-premise software companies tend to be at least 7-8 years old, with their underlying technology, tools and approach even older. This antiquated technology means they’re unable to take advantage of the tools and languages to scale across hundreds or thousands of highly ephemeral cloud computing instances and take advantage of the latest Big Data principles that have evolved significantly in the past two years.
- Engineering process and priorities. The process followed and pace at which software updates are prioritized, developed, tested and released is drastically different between on-premise and cloud offerings. To do it effectively you really need to invest in two separate teams with separate charters and visions. Most of today’s Cloud offerings seamlessly provide software updates every week, if not every day. Running two different processes and multiple versions can be very distracting and expensive, so you naturally lose focus on the important things, like operating, securing and scaling the cloud service.
- Business model and revenue recognition. Almost all on-premise software companies rely on closing deals each quarter to drive the majority of their software revenue for that quarter (perpetual licenses). Cloud companies typically take their revenue ratably over the term of the contract (subscription). Switching to a Cloud offering can wreak havoc on the financials if you previously relied on perpetual license sales.
- Sales model and compensation. Sales reps get paid on quota (how much they sell each year). On-premise deals (perpetual) tend to be at least 2-3x larger than subscription deals because all of the software is sold up front and then you pay maintenance over time. Most on-premise software companies do not properly incent sales reps to sell these cloud deals because they are not willing to pay/cost structure can't support a higher commission rate for deals that by nature are smaller.
2. Is Splunk’s cloud offering meant for the mid-market or enterprise? Though Splunk’s announcement is exciting and a validation for cloud solutions like Sumo Logic’s, I’m not sure we will actually see it in most large enterprise accounts.
- What happen to tackling Big Data? As InformationWeek points out here “Splunk Storm is decidedly not a big data play”. According to Splunk’s Web site, the pricing for advertised data plans tops out at 1TB of data. That is equivalent to less than 35GB of data per day retained for 30 days. The majority of large enterprises have far more than 35GB of machine data being generated each day – so its not clear if those customers have to move to the on-premise version or if they can scale their data beyond that. Sumo Logic is all about scale and Big Data. The cloud architecture gives us the flexibility to elastically scale any portion of our compute and storage engines on-demand, thus overcoming the headaches and performance bottlenecks of on-premise deployments.
- What about private clouds? The announcement also states that this “is for organizations that develop and run applications in the public cloud”. Splunk on their recent earnings call wen't on to say "It doesn’t have the features of Splunk Enterprise. It’s very targeted toward developers and being able to help log apps that are in development in the cloud." The majority of large enterprise applications are in private clouds, not in the public cloud, and certainly the bulk of the machine data being generated by these applications is not in the public cloud. So it would seem Splunk will be asking the majority of large enterprises to continue to install, manage and scale Splunk’s on-premise offering within their data center(s) and use the cloud offering for public cloud applications.
We at Sumo Logic look forward to seeing how Splunk evolves this first version of their cloud solution and we welcome the opportunity to address large and small enterprises’ machine data search and analytics initiatives with one highly scalable offering for public and private clouds.
Complete visibility for DevSecOps
Reduce downtime and move from reactive to proactive monitoring.