The March 2018 Cloud Native London Meetup featured an ‘Ask Me Anything’ panel discussion on Cloud Native Storage followed by a talk on Zenko, a MultiCloud Data Controller, and an introduction to Kubernetes Networking.
Alex Chircop of StorageOS (now Ondat) (@chira001) and Luis Pabón of Portworx (@_lpabon_) sat down for a lively moderated discussion on cloud native storage, followed by a number of audience questions. Subjects discussed included:
- Examples of where stateful services are needed, and gotchas that people might not have thought of when considering storage.
- What CSI is and how it affects container and Kubernetes users.
- Whether cloud native automatically implies distributed, and how to mitigate against network partitioning.
- CSI – why node level rather than container level?
- What is Rook, the newest CNCF project, and how it compares to StorageOS.
- How StorageOS (now Ondat) handles the issue of firelocking data.
Watch the Ask Me Anything on Cloud Native Storage recording.
- Zenko empowers people who create value with data – Zenko aims at allowing enterprises and individual to set themselves free from vendor lock-in, while saving them the troubles of juggling multiple APIs; Zenko itself doesn’t lock you in: everything goes in the cloud(s) in native format, so you can read from any tool; finally, Zenko simulates a single namespace over all clouds, thus allowing the user to replicate and do metadata search seamlessly across all backends;
- Zenko is open source: free to use, documented, and with an active community ready to support you! If you want to join zenkos, contribute support for a new backend, or add a microservice to its stack, get in touch on via its forum. Zenko also has an Enterprise Edition which can be stretched across sites, has a more resilient architecture, and lets you leverage the full power of AWS IAM across all clouds.
- Zenko Orbit is a Multicloud Storage Management UI that makes it easier to use its multicloud data controller; Orbit is a Pay-as-you-go service, and the first terabyte of managed data is free, so you have enough time to decide whether it’s the right fit for you or not.
- Kubernetes uses DNS for service discovery: no special client code required.
- You can’t ping a Kubernetes service, because its IP is virtual.
- Choice of pod network will likely be governed by how you get on with the people that run your host network.
Cloud Native London – April
We’re back on 3 April with more inspiring talks by speakers from Contino, Aqua, and Humio. You won’t want to miss it! Save the dates to your calendars now and register!
If you’re thinking of speaking or sponsoring a meetup, please get in touch via this form.