Most of us at Powerupcloud, are voracious readers. Since we have a strong consulting practice on all major public clouds - AWS, Azure and GCP, it is necessary for us to be in learning mode all the time because the pace of innovation from Amazon, Microsoft and Google is just breathtaking. We also have a product division that is currently building awesome android apps and executing an interesting IoT idea and a strong Analytics/ML consulting practice. So yeah, we do a lot of reading. Feedly, Reddit, Medium, Twitter, Flipboard, Quora and a bunch of newsletters. We used to be great fans of GetPrismatic for the quality content it used to surface in Cloud computing, DevOps, ML etc.
When I say quality content, it's not just news related to AWS or ML. Prismatic used to show deeply technical articles and it used to get better with more usage. Unfortunately, they recently shut the news app down just like Zite did. Looks like running news apps is not a cakewalk even if you have VC funded money. So coming back to the point of this post - we thought it would be a great idea to publish a curated list of interesting articles that we read every weekend. There are hundreds of newsletters out there and the world has more content than we can consume. But it all comes down to signal and noise. What is a signal to me can be a noise to someone else.
So who might find this weekly interesting? If you are into Cloud Computing, DevOps, Databases, Bigdata/Machine Learning and IoT - you may find some interesting content. Also some random geek stuff. With the motive thing out of the way, the next big decision was to decide what to call this weekly thing. As XKCD rightly noted, it's difficult to name things :) - So we just decided to call it by a very unimaginative name - "Powerupcloud Weekly".
So here goes the first installment. Things worth reading from last week. Well, not strictly from last week alone. It just means that we found these last week :)
Amazon Web Services
- Setting up DNS Resolution between On-premise and AWS: If you have an on-premise datacenter along with presence on AWS, you are probably using either DirectConnect or a secure VPN tunnel to connect these environments. How about resolving DNS across these multiple environments? One way is to run your own DNS servers on AWS and integrate with on-prem DNS systems. This article explains using Simple AD to forward requests to Route 53 hosted private zone. This sounds interesting but as a commentator noted on that thread - a DNS name resolution request will have to traverse to on-prem to resolve an AWS name. It seems the functionality of querying Route53 hosted private zones outside the VPC is in the works.
- DNS Resoultion between AWS and On-premises using Microsoft Active Directory: This is exactly like the above link, except we use Microsoft's active directory. Amazon has recently announced a full-fledged managed active directory on AWS, named Directory Service. This article uses directory services to configure a DNS forwarder to send requests from on-prem networks to VPC provided DNS. If you are a windows shop, you might want to try this.
- Rate based blocking with AWS WAF and Lambda: An article that explains detecting HTTP flood kind of DDoS attacks on the fly using lambda and then blocks the requests through AWS WAF before the requests hit the web server. Very useful if your are using CloudFront.
Coded Infrastructure: This is an interesting take on treating your infrastructure as code. Watch out for a series of posts on how they used AWS Ruby's SDK to achieve their vision.
- Is your code ready for the leap year?: Though this shows up under Azure category, it is a useful article irrespective of Azure. 2016 is a leap year and it is interesting to find that Azure suffered a leap year outage in 2012 and 2008 saw bricked Zune devices due to leap year. Check the leap year bugs and make sure you don't run into them.
- Azure Stack Preview: Azure stack preview is now available for download! If people don't come to Azure, Azure will go to people :). The fabric that runs Azure can now be used to manage your datacenters. If you already have workloads on Azure and continue to have on-prem, give this a shot.
Google Cloud Platform
- Global Loadbalancing on GCE: Google Compute Engine's loadbalancer is quite powerful. This is from a series of articles written by Grace on Medium about A to Z of Google Cloud.
- Loadtesting: This is a great article from GCE documentation explaining an effective loadtesting strategy using a JMeter cluster. You can use your laptop as JMeter client and send requests. During the test, the results from all nodes are sent back to the client.
- How to Replace SAP BW with BigQuery: The title says it all. Go read it. Since we are talking BigQuery, check this list of some interesting BigQueries here. Someone loaded 1.7 billion reddit comments on BigQueries and ran interesting queries. Isn't that awesome?
- Google Cloud Functions: Yep. Google's version of AWS Lambda. This space is heating up. Cloud functions are still in Lambda and currently only support
- Scaling Elastic Search: Sharding and availability for hundreds of millions of documents: This is a great article about what it takes to scale ealsticsearch. Has great explanations about how to arrive at number of shards and how to reshard with zero downtime. Must read.
- Monitoring ElasticSearch at Scale: This is again from signalfx. How do you monitor and keep an implementation of scale mentioned above.
- Scaling MongoDB to one million connections and beyond: Now that we are talking scale, how about scaling MongoDB to a million client connections. It has great talk and text of QA.
- Query mysteries: Do you have a
SQL Serverquery that runs blazing fast on SQL Server management studio but chokes when run from the application? You might be a victim of parameter sniffing. In order for you to troubleshoot, you must first understand how SQL Server's brain works. Great article.
- Postgres Query Plan visualizer: The name says it all!
That's all for this week! The future weekly posts are not likely to be as long as this one. I was sitting on a lot of links and I had to let them out in the first edition.
Happy weekend reading! :)