Views, News & more
You’ve completed your cloud migration and everything is running smoothly. However, this isn’t the time to crack open the Champagne and put your feet up. Now is when the hard work really starts!
In order to ensure that your decision to move to the cloud continues to offer good value for money and great performance, your team will need to monitor it, maintain it and optimise it.
It’s a good idea to establish who within your team will be responsible for this from the offset. Without effective optimisation, cost savings and performance levels will quickly deteriorate.
Cloud Service Providers, such as AWS and Azure, equip you with free migration tools to help with the initial deployment to the cloud. However, it’s vital that you then use the free service monitoring tools they provide to help protect against rising costs post-migration.
For Azure users the CloudDyn monitoring tool is offered until the end of 2018. It is also available as a pay-for service for monitoring Google Cloud or AWS platforms. It’s worth considering utilising this tool as it can help you make some serious cost savings. Although the cost, for use on Azure rivals, is 1 percent of annual cloud spend, the service actually shows you your utilisation levels and can demonstrate, recommend and facilitate changes that could save you money. For example, it may point out to you that downgrading one VM to a lower specification could lead to cost savings of hundreds of pounds each year. Makes sense doesn’t it.
Whether you need access to an independent monitoring tool is a question you will need to consider. You may have access to free monitoring options, but the value of a third-party service should not be overlooked. Gaining an independent insight into your ping times/uptime etc. and full visibility over your entire IT estate can be advantageous when attempting to optimise your cloud.
There are a few ways in which you can optimise the performance of your cloud, post-migration. Here, we look at a few of the simple tools you can deploy to help protect against rising costs and performance deterioration.
It’s relatively easy to add load balancers within a cloud environment. They can help you to delay upgrading of your VM specs by helping to spread loads more evenly among your front end servers.
Using your cloud provider’s database services is another way in which you can delay upgrading your VMs. Consider the database services that your cloud provider offers, as an alternative to running your own within a VM. If you are an Azure user you can switch to an Azure SQL Database from an SQL Server, or an Azure SQL Database Managed Instance, instead of a SQL Server running within a VM.
Similarly, Azure offers Azure Database for MySQL as a cloudified replacement for MySQL, and Azure Database for PostgreSQL as a cloudified replacement for PostgreSQL. It also offers Cosmos DB as a replacement for noSQL servers like MongoDB and Apache Cassandra.
There are very few reasons to host your own email unless you are an organisation of a very significant size. Opt instead for Google’s G Suite or Microsoft’s Office 365 options, which are scalable and enterprise grade.
Some app usage can lead to extremely costly licenses. If you find that you have a particularly expensive or resource-hungry app, it’s time to consider your options. Firstly, look at whether a SaaS is available to deliver the same IT outcome for which you have been using the app. Things move pretty quickly and it’s often the case that a new SaaS solution might have been developed since you first selected the app.
Alternatively, it might be more optimal to replace the app or even rewrite the app from scratch as opposed to modifying the code. This might actually be cheaper or more efficient as it could allow you to switch to more popular tech or code than the legacy choice, for example.
This is another place where costs and performance can be optimised. It is also another area where cloud providers can offer cheaper alternatives that will help to optimise your cloud performance and keep costs down.
If you are archiving masses of data, look at cloud service provider options, which are cheap, but do mean slower retrieval times. AWS, for example, offers S3 and Glacier storage solutions, while Azure offers a range of storage options for different purposes. It’s time to consider whether you actually need to be able to retrieve your data quickly - slower is cheaper after all.
Monitoring the performance of your cloud post-migration is primarily about identifying capacity constraints and cost escalations quickly so that you can fix issues before your finance department start panicking or users notice problems. Once these issues are identified, the solution often lies in changing VM specifications. Either that or switching to a completely new VM that is better optimised for the role it is fulfilling in terms of the storage or memory.
Don’t strive for 100 percent utilisation. If you are using 100 percent of your cloud capacity, you need to make some changes. Consider how your system will cope with peaks in demand. You need to monitor your usage and prevent 100 percent utilisation, as this simply indicates that you have a bottleneck somewhere.
There are a number of challenges involved with cloud deployment that your team needs to be aware of. They should be able to help you protect against VM sprawl, security breaches and cost escalation.
Depending on your business’s approach, you may also need to train them to use cloud provider managed databases, elastic computing, cloud provider APIs and serverless computing. All this will ensure that they can make the very best use of your new setup. Optimising the way in which your IT team uses your cloud is a fundamental part of getting the best value out of the technology.