Page 15 - FCW, June 15, 2016

Commentary|DAVID WANG
The case for cloud-based analytics
David Wang is director of product marketing at Databricks.
A combination of public cloud and open-source technologies could help agencies solve sophisticated problems without breaking the bank
By some estimates, the Obama administration will create more digi- tal records than all previous admin- istrations combined. Unsurpris- ingly, the explosion of data volume, variety and velocity in the federal government has created enormous challenges for acquiring technol- ogy and talent to meet the surging data analytics need. As a result of that “analytics gap,” many federal agencies are struggling to meet their mandated objectives.
Civilian institutions have suc- cessfully devised a solution through the powerful combination of public cloud and open-source analytics software. The proliferation of public cloud offerings has eliminated the high fixed cost and long lead time historically required to procure complex analytics infrastructure.
In addition, the public cloud is especially well suited for seasonal or unpredictable workloads — it can dynamically scale the analyt-
ics infrastructure up to meet high demand and shut down unnecessary capacity as needed to reduce costs.
At the same time, the maturity of powerful open-source technologies, such as Apache Spark, in the cloud has provided cutting-edge capa- bilities previously only available in esoteric and cost-prohibitive propri- etary software.
In the case of Spark, superior per- formance, flexibility and ease of use helped the technology rapidly gain widespread support in the devel- oper community and propelled it to become the de facto data processing
standard. Today, a plethora of mul- tinational enterprises ranging from Toyota to Bloomberg use open- source software such as Spark to solve many sophisticated problems, including monitoring sensor data and detecting fraud.
Given the establishment of public clouds dedicated to the federal government, agencies can also take advantage of the technology. Like- wise, open-source analytics soft-
The public cloud is especially well suited for seasonal or unpredictable workloads.
ware is readily available via Amazon Web Services’ government cloud offering. In short, the stage is set
for agencies to embrace the latest cloud-based data analytics technol- ogy and tackle more sophisticated problems while staying within bud- get constraints.
For those who are planning to migrate to cloud-based analytics, here are three simple best practices to keep in mind to maximize its potential:
• Embrace technologies endorsed by the developer com- munity. The best open-source projects evolve at a frenetic pace because of contributions from
the developer community. Choos-
ing software with broad support enables you to capture the benefits of rapid innovation while advertis- ing your organization as an attrac- tive place to improve one’s skills with elite peers.
• Think holistically about your needs. For most organizations, the true power of analytics stems from combining different approaches — such as real-time stream processing and machine learning — to answer more complex questions. Under- standing the different scenarios your organization might encounter in the next three to five years and choosing the technology foundation that can tackle them will allow you to future-proof your investment.
• Offload noncritical work to maximize productivity. When working with sophisticated technol- ogies, a significant amount of main- tenance is necessary to keep the infrastructure operational. However, those tasks reduce the productivity of the teams responsible for deliver- ing analytics. The cloud provides an easy solution to mitigate the burden in the form of managed services that allow routine maintenance tasks to be offloaded via automation soft- ware, enabling your team to focus on the mission.
With these best practices in mind, the powerful combination of cloud and open-source software will allow agencies to eliminate lengthy wait times to procure infrastructure, attract and retain top technical tal- ent, and meet the growing demand for analytics. n
June 15, 2016 FCW.COM 11