IBM Cloud Series – Intro
We are starting a small number of article series related to some of our experiences on IBM Cloud platform
L3C has been involved in IBM Cloud projects over the years in its determination to know more about the major public cloud players on the market. Moreover, our team has always had a particular interest in public cloud technologies and services.
We would most certainly advertise our own cloud platform, however, in specific cases IBM Cloud or a more public cloud solution would be required by our customers in their ongoing projects.
Most of our implementations have been performed around VPC and classic infrastructure environments, with a different level of security and availability involvement.
Future topics
Thus in this series, we would like to share our impressions on a few main topics:
- Security – user access, encryption, high availability, etc.
- Computing in IBM Cloud – instances, images, plugins, etc.
- Storage – object and block storage, replication, etc.
- Networking – dedicated firewalls, security groups, VLANs, subnets, etc.
- Monitoring – basic and Sysdig
- Backup and recovery – IBM Cloud Backup(eVault) and Veeam implementations
- Disaster Recovery- DR implementations in IBM Cloud – including different data centers, Global IP addresses, etc.
- Support – SLA, support levels, escalations
Changing conditions
Even if more than 7 years have passed since IBM bought Softlayer to establish its public cloud division, things still feel a bit transitional. You will very often stumble upon documentation from Softlayer times, non-existing referrals or performance limitations due to older equipment in data centres. The services are constantly changing – mostly for the best – and we apologise if things have since moved on but our story is based on our experiences during a real implementation in the second half of 2020.
Going in carefully as always
It is needless to say that even small implementations in public cloud environments like IBM Cloud should go through a stage of design, planning, testing and deployment. This is even more valid if you are going into the environment for the first time. Something that might feel light at the beginning, could create a number of issues for you subsequently. So, it is imperative to plan and test. Generally, if you work with implementation experts, they would have some credits to provide for an initial PoC.
IBM Cloud interface is rather intuitive although it needs some time to get used as there is complexity in user management, billing and notifications. If you are doing this alone and have questions, there is always support on the chat (which I often use for quick questions) or ticket system for more complex inquiries. We will explore that in а following article.
Work your way through the documentation at their docs portal and start deploying.
Hourly / monthly decisions
There is a scheme for most things on an hourly or monthly billing basis. If you plan to test quickly – for less than a month it is obvious that an hourly rate would be more suitable for you. However, keep in mind two main things:
- You will pay a minimum 25% of the cost of the server irrespective of how much you actually use
- Hourly will not allow tests for some features that are essential to test – advanced monitoring, backup, etc.
During your tests, of course, be sure to stop instances when you are not using them to save money on an hourly basis. Some things, like additional IPs and storage, will still be charged, so plan these costs as well.
What we have noticed for IBM Cloud is that documentation is scarce outside of the official channel, which is really unfortunate as the most helpful use cases are always outside of the branded portal. So, keep notes for everything you learn as you might come back to it again.
Lessons learned - sounds awful 🙂
So, some of the initial lessons learned (more will come in the other topics):
- Design and document well
- Ask everywhere you can
- Test before implement
- Configure notifications wisely as you might receive notifications that IBM Cloud is stopping your server to migrate to a different host and this to your order@… email.
- Take your time – especially networks need some getting used to even for a network expert
- There are some API functionalities, which can really save you time if you deploy on a larger scale
- Monitor everything, you will not regret it
More will follow
OK, that’s enough for a starter, we will keep more content coming on this later on.
You can learn more about our IBM Cloud Service here.
If you would like to contact us with some questions or a project, we would be happy to help.
Note: L3C is not aligned to any particular public cloud provider and is an independent cloud services provider. This blog is solely to share our experience and is based on real implementations.
