In-memory computing may be all the rage in the technology world, but what is its use for a small business, who may not have that much data they need to crunch on a regular basis? According to Sharma, “It's not always about the technology, but the business outcome you can derive from it!”. Here is the Q&A session we held with him:
Q: What fundamental problem does in-memory technology solve?
Sharma: In-memory technology is gaining a lot of traction predominantly because of advancement of hardware and dropping costs. According to Moore's law, cost of chips are decreasing year on year and hardware is exponentially improving. Although there have been advancements in hardware, the single point of latency has still been the disk. In-memory helps you to solve this problem by storing all the data in-memory itself. In-memory can be extremely effective when dealing with unstructured data, such as social media data.
Q: How can it be beneficial for SMEs, especially those without a lot of data?
Sharma: It purely depends on the business scenario a business wants to use it for. Even if you're a SME customer, you might have a lot of social media analytics that you may want to analyze. For example, a company may want to do mobile ad placement based on what's happening in real-time, and this is enabled by in-memory powered processing. Even if you're an SMB, you may have a mobile app for your employees to punch in their CRM data on the go. If it takes 1 minute to process and update the database at the back end, the whole user experience of the app goes away. However, if you have an app built on the fundamentals of in-memory technology at the back end, transactions will be completed in a few seconds, and this will be a huge asset to the salesforce, who is always on the go and needs to update their data from anywhere. The key is that in-memory such be seen as a combination of a way to deal with “big data” and provide a seamless user experience at the same time.
Q: How will in-memory technology affect business processes?
Sharma: The paradigm of how applications are create will totally change, just as how we are consuming data changes. For technology teams at the back end creating applications, there will be a paradigm shift in thinking as they start to optimize algorithms for use with in-memory technologies. At SAP, we have engaged with many start-ups to explore possibilities at what impact our HANA platform can have in their product offerings. Our HANA(High performance analytic appliance) platform is unique in that we do not force it to be used solely on the SAP ERP/CRM platforms. It can be made compatible with products from other vendors as well.
Q: So does that mean 'Big Data' is not a factor for in-memory adoption?
Sharma: There is no doubt that in-memory computing has become relevant because of “big data”. However, the real-time visibility into your business itself is a paradigm shift. The traditional approach of doing analytics is moving transactional data into a data warehouse,creating reports and exposing them to end users. From the time the transaction has entered inventory to the time the information is exposed to the user, it might take hours or even days! With in-memory, you can even store large amounts of historic data so reports can be generated almost instantaneously! Even for SMEs without a lot of data, in-memory offers a standardized data management that reduces complexity of moving data around and offers it in real-time, which can be key to decision-making.
Q: How does a business implement in-memory technologies?
Sharma: An in-memory solution will be a mix of both hardware and software. For example SAP's HANA comes as an appliance. It is based on intel x86 servers, so it comes with its own chipsets and RAM preconfigured. In-memory databases have capabilities to do number-crunching and data persistence embedded in them. In-memory technology is just a technology if there is no business outcome intended with it. SAP talks to many businesses and challenges them with “What If” scenarios to identify the impact it will have in the supply chain of the business. Only if the business has identified a clear business goal of in-memory will it be of value to them. There are also software-based in-memory solutions available, which leverage the hardware they are installed on, and bring their own techniques of delivering in-memory technologies, such as the form of a live cache. They can even leverage on other hardware available in the network to perform distributed processing.
Q: What challenges do SMEs face in implementing in-memory solutions?
Sharma: Small businesses must realize that migrating to in-memory is not done overnight. There has to be an incremental change and adoption to in-memory. Moving their data from traditional database to an in-memory database is challenging, but there are solutions to this. Businesses can have a side-by-side architecture, so that data is being replicated into the in-memory database as soon as the transaction happens. For example, HANA has something called a replication server which takes data transaction by transaction and replaces it into the HANA database.
Q: What kind of industries benefit the most from in-memory technology?
Sharma: It will be very beneficial for point-of-sale retailers and loyalty program distributors. For example, a store may be having a standard 15% discount for an item. However, if they have a loyal customer who has been around for a long time, they expect that they will receive a better offer. Real-time offers such as this need to run through a lot of data, which can only be done quickly using in-memory computing.
Q: What is SAP's offering for small businesses? And how does it differentiate itself?
Sharma: We have a business suite called Business 1, which is catered towards customers in the SME segment. We offer SAP HANA in this suite as well, which provides an in-memory solution suitable for smaller businesses. Additionally, customers can also deploy applications onto HANA which are not ERP, so it's not specific to ERP applications. All of SAP's applications will soon be migrated onto the HANA platform. We believe that other vendors have a mix and match approach, but since SAP has no legacy in this field, we have built an architecture from the ground up. SAP also has engagement with customers across 23 industries to discuss how our products can fit into their product offering. It's not about the technology, but the business outcome you can derive from it!