What it Means to be Cloudera Certified (Part II)

Cloudera Certified

In Part 1 of this blog we talked about how the Cloudera Connect partner certification program is the most comprehensive in the Hadoop market.  We’ve achieved a level of completeness in our certification program that brings value for all participants. Our platform stack is as deep as it is wide, and it is covered with one all-encompassing process that captures the nuances of every integration. We certify partner products across the breadth of our stack, including but not limited to the following integration points. The categories below and allocations of components to the categories is flexible and somewhat loose, but this classification scheme is useful for establishing a common language when talking about our partner ecosystem.

Ingest / Storage

  • Hadoop CLI
  • Flume
  • Sqoop
  • Kafka
  • Avro / Parquet formats
Batch Processing / Analytics

  • MapReduce
  • Hadoop Streaming
  • Pig
  • Hive
  • Kite SDK
  • Spark
Query / Interact

  • Impala
  • HBase
  • Accumulo
  • Cloudera Search
  • Hue
Infrastructure / Metadata

  • Cloudera Manager parcels
  • Cloudera Manager  API
  • Custom CSDs
  • Cloudera Navigator
  • HCatalog
  • YARN
Machine Learning

  • Oryx
  • Mahout

In order to reach certification status, the certification test of the partner product must exercise each integration point and must also show that the use of that integration point doesn’t preclude others. For most partner products, this goes without saying and is easy to achieve, but there are opportunities for joint business at every level. The deeper you go down the stack, the more deliberate we must be in certification as the disruptive potential is higher. We currently classify partner products as follows, with the darker colored background denoting a more involved integration and a more rigorous certification:

On the top end of the range you might see a BI vendor which offers dashboards and visualizations via ODBC to Impala, essentially treating it as yet another database. On the bottom of the range you’ll see a specialized hardware appliance that comes bundled with Cloudera software, or a pre-built Hadoop solution built on industry-standard hardware. Near the bottom, you’ll see a product that actually offers an HDFS replacement to round out an EDH offering and touches our entire stack.

The vast majority of partner products we work with are somewhere in the middle, offering Hadoop-specific functionality above and beyond simple SQL, but also leveraging the benefits of our distribution.

This granularity of classification allows us to cover all players in our partner ecosystem with the specificity they deserve, whether they are a Hadoop-native player or a long established organization. It also allows us to answer detailed questions about the specific nature of every certified integration.

Cloud, hardware, and infrastructure partners are the most tightly integrated and most involved. We have a dedicated team to this segment, devoted to working closely with these partners on joint reference architectures and validations. No partner in this space obtains certification without working with this team, and no customer purchasing a joint solution should deploy on an architecture that strays too far from a Cloudera-blessed RA.

Automating the Certification Process

On initial explanation, the depth and breadth of our platform and the partner ecosystem it touches may look daunting, but the maturity of our product and our business makes certifications at this granularity manageable in a scalable way.

We leverage the diagnostic bundle capability of Cloudera Manager as well as the analytic capabilities of our internal EDH to automate the validation of the quantitative aspects of the partner certification tests, such as making sure the certification environment meets the security requirements and capturing the integration points exercised during the test. This frees our Partner Enablement Certification Team to focus on the qualitative work of building partner relationships, recording partner demos for our field and support personnel, engaging in Q&A or support activities for our partners, and helping our customers through their joint deployments.

Being rigorous in our certification requirements does not necessarily mean that the level of effort required from the partner is higher. That’s because Cloudera Manager makes cluster setup and configuration easier than ever before. Cloudera Manager 5 almost completely automates Kerberos setup. Since single-node VMs are not allowed for certification, partners can use Cloudera Director to spin up robust clusters in the cloud of significant size for short periods of time, allowing for a complete certification test while keeping their investment to a minimum. Cloudera offers access to complementary certification environments we create using Director. In those cases, all the partner has to do is take the keys and get certifying.

As our EDH platform expands to tackle board-level, business-defining use cases, the partner certification process has never been more important. Our customers should expect any partner that carries the “Cloudera Certified” logo has been deployed in a real environment that meets security requirements on a realistic dataset.

We love to hear from our customers. Let us know how our certified partners are working for you.

Interested in certifying with Cloudera? Join the Cloudera Connect partner program for more information.

The post What it Means to be Cloudera Certified (Part II) appeared first on Cloudera VISION.

Leave a Comment

Your email address will not be published. Required fields are marked *