Since the time Amazon announced its Database-as-a-Service, most IT vendors, big and small, have either spun off services utilizing Amazon or built services that are complementary to Amazon. And many others have rolled out competing products. Each of them is targeting SMBs with the promise of agility, reliability, scalability and integration capability. The market has become crowded with options that are very difficult to sift through for SMBs. Nevertheless, the interest in database-as-a-service has never been higher. As per Techaisle’s 2013 SMB Cloud Computing study, between 6% and 64% (varies by employee sizes) of SMBs are either using or planning to use database-as-a service. However with an average number of formal IT staff at 4.2 and percentage of SMBs having formal IT staff varying from a low of 3% (for 1-4 employee size) to high of 97% (for 500-999 employee size) businesses the adoption is sporadic and hindered because of inherent possibility of “lights-out” situation due to outages, network latency and connectivity issues.
GenieDB, a company founded in 2011, has brought into the market a MySQL-database-as-a-service that that removes outages from the equation. Phyken Media, a video game studio for mobile platforms, developer of “Wizard Ops Tactics” is GenieDB’s marquee customer. Kunal Patel, President of Phyken Media, was faced with two huge challenges;
1/ during development of the game he had to have access to robust technology that could scale to production rapidly,
2/ be able to deploy into multiple geo locations for multiple-cross-region-player challenges without having to install and manage multiple servers.
In such a rapid-fire environment, where all the data is “hot”, Kunal and his team of developer-artists turned to GenieDB to utilize its MySQL database-as-a-service offering which he said is hugely disruptive enabling globally distributed databases where all nodes remain synchronized.
To get to know more about GenieDB we had a very detailed Q&A with Sumeet Sheokand, CTO, GenieDB. Given below is an excerpt from the interview.
Techaisle: What is the motivation behind GenieDB and in particular MySQL-as-a-Service?
Sumeet Sheokand, CTO: GenieDB is built with the aim of dramatically simplifying database management in the cloud for business critical applications. Dealing with cloud outages, network latency, redundancy, replication, tuning, etc. can be very painful and time consuming, distracting precious IT resources away from other core business areas. We want to evolve this vision to the point where you no longer need a team of DBAs to manage complex, distributed database platforms. We will either automate processes or provide a very simple interface to manage hundreds of highly available, low latency MySQL databases ready for any cloud infrastructure, anywhere in the world. It is well known that the only way of overcoming the challenges of cloud outages and network latency is to distribute copies of the database across wide geographical areas, a historically difficult problem for relational database architectures. GenieDB makes this critical functionality available to all businesses and all budget sizes, in a very easy-to-use package. In keeping with the theme of making it easy, we offer our core technology as a Database as a Service (DBaaS). This allows us to offer our users a point and click ability to deploy geographically distributed database servers with automated management and monitoring already in place.
Techaisle: What problem are you trying to solve? Especially, how can small and mid-market businesses benefit from GenieDB?
Sumeet Sheokand, CTO: We have created a geographically distributed database fabric that removes the database as a single point of failure and makes cloud-enabled MySQL database provisioning, management and monitoring an exceedingly simple affair. With our service, the small and mid-market businesses can focus on running their businesses or building their applications rather than worrying about architecting complex distribution, replication and failover systems, not to mention installing, patching and maintaining backups. All the rote tasks of using databases will be taken care of by our service.
Techaisle: Is the service a better mousetrap than others that are available in the market?
Sumeet Sheokand, CTO: The comparison between what is available in the marketplace today and GenieDB can best be described as the difference between a database-in-the-cloud vs. a Cloud Database. The existing DBaaS solutions are simply databases which have been put into the cloud utilizing a variety of scripts (i.e. "A Few Scripts-as-a-Service"). These offerings are partial solutions as they do not remove the complexity of building distributed systems in the cloud. Some only offer single locations; some only offer it in certain locations, while others don’t offer an easy path to grow the database with business traffic. Through our proprietary storage engine, GenieDB has fundamentally transformed the way a relational database functions in the cloud, removing all of these common pain points within the database layer. Businesses can choose their cloud providers, their locations and size, and know that their database will be available, responsive and grow as their business grows, with minimal effort.
Techaisle: How do you think it is different from Translattice, NuoDB, Galera?
Sumeet Sheokand, CTO: GenieDB stays true to MySQL, the most commonly used database in the world, as compared to Translattice (Postgres) or NuoDB (Not MySQL, custom). GenieDB also offers an ‘Eventually Consistent’ model that allows us to place the nodes as far apart around the world as the business needs and deliver local database performance, instead of being limited by network distance (Galera).
Techaisle: Is GenieDB’s approach similar to Google’s Spanner?
Sumeet Sheokand, CTO: GenieDB has some conceptual similarities to Google Spanner, such as a consistent Clock around the cluster. GenieDB uses software Lamport Timestamp based clock compared to Google’s hardware based clock. Even though there are similarities, GenieDB was developed independent of the Spanner effort. It was nonetheless a great validation of our vision and approach. Google Spanner is also not available as a product for businesses at large to use in their own applications. GenieDB is available today to reap the same benefits.
Techaisle: Can it be deployed only across Amazon or other cloud providers as well? Do end-users have a choice of cloud provider they feel comfortable with?
Sumeet Sheokand, CTO: GenieDB is currently available across Amazon, Rackspace and Google’s cloud services. We are actively working on integrating a handful of other popular providers, including HP, into the service. We will be happy to work with our customers to accelerate their cloud of choice for integration.
Techaisle: The future is Hybrid as we all know it, so how does it work if in a customer’s environment there are both on-premise and cloud based MySQL databases?
Sumeet Sheokand, CTO: GenieDB core software is agnostic to machine location. It just needs a Linux machine with network access. It is because of this capability, that we can build a database cluster across multiple cloud providers. Hence, we can support any Hybrid configuration, if we have access to the machines. There are business and security concerns on providing full access that would need to be figured out.
Techaisle: Why did you choose MySQL for database-as-a-service?
Sumeet Sheokand, CTO: MySQL is the ubiquitous database for modern applications. It is by far the most widely installed and used database in the cloud today and hence is the basis of our service.
Techaisle: Most of the SMBs do not have IT staff let alone database administrators. Can it still be deployed with limited technology knowledge?
Sumeet Sheokand, CTO: Deploying GenieDB does not need any technical knowledge, staff or application changes besides pointing the application to the GenieDB provided database location. It is this [SMB] specific target audience for which we designed GenieDB.
Techaisle: How do you help SMBs in understanding the technology and how the service solves their business pain points?
Sumeet Sheokand, CTO: GenieDB has extensive amount of information available at its site including FAQs, White Papers and Demonstrations. We also offer a one week, free trial of the service and support so that potential customers can experience the service before buying. We are of course happy to spend as much time as necessary to get any new client comfortable with our service.
Techaisle: For type of SMB customer is this solution most suitable?
Sumeet Sheokand, CTO: GenieDB is a Storage Engine for MySQL and as such is a general solution. Hence, GenieDB works with any application that works with MySQL today. From a business perspective, any business that is running a critical application on a single database could benefit from GenieDB, so that if one database server goes down, others are still available and the application will continue to work without any downtime. Typically we see customers that are starting a new project or that have an existing project that is anticipating growing traffic or is becoming more business critical.
Techaisle: What is a typical deployment timeframe?
Sumeet Sheokand, CTO: Nodes are spun up and the cluster made available within minutes. Most applications can be up and running against GenieDB under an hour, including data transfer. The actual duration does depend on the amount of data to transfer and the upload capacity available from the customer site.
Techaisle Take
At last count, MySQL is still the most widely used database for cloud deployments and is easily the one that SMBs use extensively. However, MySQL’s continued dominance is being questioned since it was ingested by Oracle through its Sun acquisition. Although Postgres is usually considered to be more powerful and meant for big data sets, there is yet no visible mass migration from MySQL to Postgres. The success of GenieDB and the decision to use its DBaaS is not dependent upon the debate of MySQL vs. Postgres but GenieDB’s relevance for the most-used open source database. GenieDB has created a very useful solution for SMBs but will be challenged to capture mind-share of IT consultants, service providers and the developers who are advisors to SMBs for development of applications based on multi-nodal, geographically dispersed, tightly synchronized MySQL databases. Apart from focusing on growing its customer base GenieDB will have to continue to add feature sets, capabilities and integrate with different regional cloud providers. For now, GenieDB’s MySQL-as-a-Service receives check marks for its simplified usage, ability to integrate with all MySQL applications, rapid deployment, cloud portability (not being locked-in with a single provider) and above all making the database immune to outages.
Techaisle Analyst Insights
Where is the market for client devices going – and what does it mean to corporate strategy? These are questions that Techaisle and its clients – and the industry as a whole – has been wrestling with ever since the decline of the traditional PC opened the door for alternative client technologies.
In many cases, users now combine mobile and fixed devices, and with this multi-device approach, consistency and security become even more important than they were in the laptop era. Many organizations are responding to a need to securely manage and distribute user data and applications by investigating virtual desktop (VDI) technologies enabling delivering of “desktop as a service,” or DaaS. With VDI/DaaS, businesses deploy client virtualization technologies from suppliers like Citrix and VMware to ensure that users have anywhere/any time/any device access to current information, their applications and their desktops. These technologies allow for better data security and auditability, and often offer the additional benefits of reduced CAPEX and OPEX.
The allure of VDI and DaaS is clear – but the technology itself and the path to realizing its benefits can still seem somewhat mysterious to many small and mid-sized businesses. To understand implementation challenges and lessons learned, Techaisle conducted depth interviews with small and mid-market businesses (from 50 employees to 999 employees) that are currently using VDI and/or DaaS solutions. Based on a random sampling these businesses belonged to financial services, professional services, manufacturing, construction, utilities, retail and private education segments. These early adopter SMBs have rolled out VDI and/or DaaS solutions within their organizations. The number of users using VDI and/or DaaS within the businesses interviewed ranged from a low of 30 employees for a small business to a high of 600 employees within a mid-market business.
Techaisle’s quantitative VDI/DaaS research shows that the key user objectives in adopting either on-premise or hosted VDI/DaaS solutions revolved around mobility, application availability from anywhere and on any device, disaster recovery, centralized management and administration of end-point devices at the same time reducing costs. Based on users’ real-world experiences and feedback, Techaisle has compiled a list of seven key lessons for success for small and mid-market businesses planning to adopt VDI/DaaS solutions.
1. Prepare a roadmap of the solution and a blueprint of implementation process
Before starting the implementation process and before even considering a pilot, current VDI/DaaS users advise potential users to create a roadmap of the solution and a blueprint of implementation process. The roadmap and blueprints should include solution and brand selection criteria, a list of solution components and their objectives, changes required (if any) to the core infrastructure to support the solution, costs involved and budgets for cost overruns, security vulnerabilities, phase-in of users and their training process and timeframe for enhancements post implementation.
2. Hire external consultants - people who have experience
It is almost impossible to develop a roadmap and blueprint without the guidance and close involvement of experts. All SMBs that we interviewed had contracted with external consultants varying in size from a group of 4-5 to a maximum of 40-50 people. In each case, consultants were preferred over resellers and service providers due to their focused deep expertise and track record. However, each of the current users of VDI/DaaS had decided on either Citrix or VMware solution prior to engaging with consultants with corresponding expertise.
“We approached our known consultant which is a small company formed by 6 to 7 people and have expertise in VMware solution. They are highly technical professionals providing free infrastructure and free connectivity support and covering infrastructure maintenance and end user connectivity”.
“We preferred going with Consultants as they had solution specific expertise. They gave optimal options taking into consideration both cost and technology sound solution. They even offered technical assistance even after post implementation and maintenance for one year. They are a small company with VMware expertise formed by a core group of people”.
“We approached consultants, a small group of people, with deep Citrix expertise, because they had relevant qualifications and certification”.
3. Ensure that the solution supports legacy software
Not all SMBs are using all modern applications; many mid-market businesses have core legacy applications and/or applications that are essential to a specific department and workgroup. Current SMB users of VDI/DaaS advise that the blueprint prepared should include a list of applications currently being used within the organization and to systematically test to ensure that they will continue to be supported in the new DaaS environment.
The VDI/DaaS SMB users also advise that businesses should also revisit the current licensing arrangement they have for some of their applications and ensure that in a virtualized environment those licenses are valid and applicable.
“We learned during the implementation phase that not every application is supported by VDI or DaaS solution. It is difficult to understand and analyze the amount of storage used by different departments”.
“The main concern was with software licensing concerns. Few of our application’ licenses restricted the use of software on systems accessed by terminal servers. There were issues coming up initially that affected our end users as these applications were not accessible on multiple virtual desktops”.
“The main concerns were the length of the project, the cost of the project and back end integration. Backend integration was a major concern as we have legacy application running on the system”.
4. Get the business users ready
Business users do not like changes that affect their interfaces – it takes time to build new usage habits, and this can (and generally does) have a short-term impact on productivity. Although a VDI/DaaS roll-out is often used to deliver better mobile systems to business users – generally, a well-received benefit – it is essential to prepare the business users with new interfaces, log-ins, support mechanism, and training on the use of thin clients. This will require IT to manage a number of VDI/DaaS-specific issues, including application downloads to user devices, management of persistent or non-persistent desktop experiences, and tactics to address latency if/where it impacts performance.
“A challenge we faced post implementation was to manage new rounds of user trainings. The new solution meant introduction of a different system to the users and one where users had to undergo a series of trainings to become comfortable with it”.
“The main challenge that we faced was user training and that was a really big concern for us”.
“We had to give training to users in about two weeks which was taken out of their operational hours and once the service was put in place the learning was steep, the users were not very productive during that time”.
“Most of the issues were to do with the users who were unable to come to terms with the changes and the content that could access easily”.
5. Conduct a pilot
Conducting a pilot helps in fine tuning the roadmap and blueprint for implementation.
“During the pilot test, we observed a need to modify our terminal server as they were not responding to the end users request. Then we decided and made changes in terminal server by making a cluster of terminal servers so that similar end users request would be sent to desired terminal server, to cut down network traffic congestion”.
“At the initial stage we started facing issues which were basically related with bandwidth or poor network response. We decided to increase our bandwidth for offering end-users customizable experience similar to that of a physical desktop”.
“In the pilot stage some issues popped out. First of all, the expected cost for the pilot stage rose considerably. Also we got a mixed reaction from the team using it as some said they were able to fully utilize the resources, whereas when we tested it over other networks like WLAN, the data was not accessible”.
6. Create a detailed budget, be prepared for cost overruns
The current users of VDI/DaaS suggest that potential adopters should budget 25 percent for software, 20 percent for services, 20 percent for networking and one-third for hardware. The percentages vary for small businesses where the budget allocation for hardware varies between 10-15 percent and the proportional cost of software rises to 30-35 percent. Current users advise businesses that have legacy applications to allocate higher budgets for hardware and services, as high as 45 percent and 35 percent respectively.
As many as 40 percent of SMBs indicate that cost overruns of 10-20 percent is a given.
“There was additional expenditure required. The testing phase was difficult as we had to change our server and switches”.
“In testing phase we found out that the users had to get software assurance which delayed the project, delayed purchase of licenses and forced additional expenditure”.
7. Upgrade server and network infrastructure
The most common and almost universal changes to the IT infrastructure to prepare for VDI/DaaS usage are installation or upgrades of blade servers, upgrading cabling to fiber optics cable thus enabling substantially higher data bandwidth, replacement of switches for routing higher throughputs, installation of thin clients and in some cases increasing storage.
Current users generally opted for specific server and thin client brands based on either recommendation from their consultants or because of existing relationships. Brands most often used were IBM, HP and Dell. Most SMBs preferred to use Blade Servers.
“We upgraded cabling by using fiber optics cable which boosted our bandwidth and smooth flow of data from the data center to the end users. We used fiber optics as it was a cost effective solution for us rather than going on with existing metal cabling which had an impact on bandwidth”.
“Networking and replacement of cabling was one issue as it did not work with the solution initially, the cabling between the server switches and office switches and for this we looked for fiber cabling”.
Concluding observations
The need for VDI/DaaS as a mobility enabling technology is clear, and its corresponding benefits for user experience and data management are compelling. However, the path to VDI/DaaS can be tricky to navigate. By capitalizing on the advice provided by current users, SMBs interested in adopting VDI/DaaS can set realistic objectives and expectations, and can manage confidently towards effective deployment.
Linoleum vs. Deep Carpet Selling
There is an old story about a consultant who was advising a client about changes in his market, and what they would mean to sales strategy. The consultant went through a series of tables demonstrating that, due to increased interconnectivity with other corporate systems, products in the client’s segment were increasingly purchased by senior managers rather than shop floor managers. The consultant stressed the importance of developing new marketing material and directing the sales force to call on the senior managers instead of the shop floor, to which the client replied, “You are talking about deep carpet selling. We don’t do deep carpet selling. We do linoleum selling here.”
Most IT vendors engage in a variation of “linoleum selling”, focused on engaging IT professionals in discussions that focus on the technical attributes of their products. However, BDMs (Business Decision Makers) – who tend to inhabit the “carpeted” realms of their businesses – are more likely to be engaged by discussions about business benefits and objectives than by “feeds and speeds”. In categories where the BDM is central to the needs identification and budget process, sales reps will need to develop “deep carpet” language and skills.
The data from Techaisle’s SMB and Midmarket IT Decision Making Authority survey demonstrates that we have already reached that point in cloud applications and software in both the small and midmarket segments.
Need vs. Enhancements
Survey data shows that in both micro/very small businesses (1-19 employees) and the smaller midmarket businesses (100-499 employees), BDMs are the primary drivers for determining the need for new cloud business applications.
These findings are broadly consistent with the results from the survey question on determining the need for enhancements to existing cloud solutions. While in most cases, IT has more influence in determining the need for enhancements than it does in determining the need for new solutions, BDMs are still generally the most important voice in the discussion. ITDM’s (IT Decision Maker) influence is directly attributable to the extent that enhancements are driven by technological rather than functional requirements.
Conclusively, survey data shows that business requirements are the prime mover for identifying the need for both new solutions and significant enhancements/upgrades in micro and very small businesses, and that technology concerns play a meaningful role in instigating discussions about enhancements to existing solutions in businesses with 20-499 employees.
Interestingly, within the 500-999 segments, there is more BDM influence over identifying the need for meaningful enhancements than for new applications. Following the logic applied to the other segments, this suggests that enhancements within these near-enterprise accounts result primarily from process optimization requirements, rather than from a need to upgrade the underlying technology.
Cloud Software budget authority
“Determining the need for” a new business application or a “meaningful enhancement” to an existing application is not, of course, identical to signing off on the purchase of a new solution. When Techaisle extended its questioning to include “budgetary control and authority,” it resulted in two interesting findings:
- The proportion of SMBs where budgetary control and purchasing authority for new applications rests entirely with BDMs increases in all employee size segments, relative to the statistics for determining need in these segments. This means that BDM control over the final purchase decision is even higher than the “determining the need for” statistics suggest.
- The proportion of respondents reporting that responsibility resides entirely with either IT or business – but is not shared between them – increases in five out of seven employee size segments (missing only the 10-19 and 20-49 employees groups). This suggests that needs identification may be more collaborative than final purchase decisions.
Both findings point to the same conclusion: that BDMs are extremely important to suppliers of cloud software. Chart below provides a graphical representation of the determining need vs. final purchase decision authority balance by employee size.
Is IT losing its authority over IT expenditures and directions? Data from the Techaisle report “The 360 on SMB & Midmarket IT Decision Making Authority” suggests that increasingly, business decision makers (BDMs) make technology-related decisions and control technology-related budgets.
The report finds that SMB “Shadow IT” in the US – expenditures made by business management without IT involvement – will amount to $27 billion in 2015. Added to the “formal” IT budget that is visible to IT but under BDM management, technology spending by US SMBs that is outside the control of the IT department will reach $99 billion, a figure that is greater than Microsoft’s annual revenue, twice the revenue of Cisco, and nearly 25 times larger than the revenue recorded by Salesforce.com in its fiscal 2014.
The data clearly illustrates that the earth has shifted from underneath the IT department within small and midmarket businesses. Executives in these companies need to understand what these new spending patterns mean to IT deployment and efficiency within their operations, while suppliers to this market – business application vendors like Microsoft and Salesforce.com, hardware vendors like HP and Dell, and the thousands of services firms that help US SMBs to make sense of technology – need to adjust to the changing patterns of SMB IT investment and control.
Shadow IT is a commonly-understood phenomenon: it represents spending on IT products and services by BDMs that are made without the IT department’s approval, guidance, or in some cases, even without IT’s knowledge. IT itself generally portrays these purchases as dangerous to the organization, creating the potential for security breaches, incompatibility between corporate systems, inconsistency in corporate systems of record, and/or loss of critical data. BDMs tend to portray them differently, positioning these purchases as IT extensions to current business activities that respond to business needs more quickly and directly than the IT department is capable of doing.
Whatever one’s perspective on shadow IT, it is clearly an important force in the SMB IT market. How important has been a matter of conjecture, since by its nature, shadow IT is difficult to isolate and quantify. However, by comparing multiple data sets from surveys that capture both ITDM and BDM perspectives, Techaisle is able to provide fact-based estimates of shadow IT activity within US SMBs. Highlights of these findings include:
Shadow IT spending on business applications
Authority for “formal” business application spending varies widely between small and midmarket businesses. However, the overall level of shadow IT spending on business applications is very consistent across the two SMB segments, at 15 percent of total small business application spending and 14 percent of midmarket business spending. In addition, business management (BDMs) within SMBs formally controls over 50 percent of business application expenditures.
Shadow IT spending on infrastructure products
The infrastructure products market is much different than the business application market – both across small and midmarket businesses and with respect to the influence of IT over “formal” purchases. The influence of IT is much greater in the infrastructure category than in business applications: IT is responsible for 23 percent of infrastructure spending within small businesses and controls well over 50 percent of total spending on infrastructure within midmarket businesses.
Overall, shadow IT accounts for 56 percent of small business infrastructure expenditures. The enormous shadow infrastructure spends by small business indicates a clear problem for small business IT managers, and realistically, for small businesses themselves: the notions that shadow IT creates security and related issues are not merely an IT construct, it is a real issue. Suppliers with solutions that help address shadow infrastructure problems (such as MDM, managed app stores, etc.) will find a very substantial potential market in the US small business segment.
Shadow IT spending on IT services
BDM-led spending on IT services has different implications in different employee size categories: in small business, it often represents an authorized or “formal” spending on mainstream IT services, while in larger businesses, it may represent a means of avoiding IT department involvement in new IT/business initiatives. Techaisle data supports this perspective. BDMs control 35 percent of IT services spend in midmarket businesses. The shadow IT spending within the midmarket – pegged by Techaisle at 48 percent of the total – creates an intriguing opportunity for IT services suppliers. “Official” suppliers to midmarket businesses may continue to sell to IT, which controls a higher proportion of the formal IT services budget than their BDM colleagues. However, when shadow IT is added into the opportunity pool, BDMs are as potent a force in the midmarket business IT services market as ITDMs. This suggests that two different approaches – positioning IT services firm as an extension to IT, or as an alternative to IT – have equivalent market opportunity today.
