Just as you’ve likely heard “there’s an app for that”, you can probably find a metric, too. But what exactly are they, and why bother with them? The digital world revolves around metrics and analytics, stats and data. Metrics are informational. Analytics are strategic.
In its simplest dictionary definition, metrics are “a standard for measuring or evaluating something, especially one that uses figures or statistics”. They collect and present data. More specifically, TechTarget defines a metric as “the measurement of a particular characteristic of a program’s performance or efficiency”.
How you use that – if at all – is up to you. And while there are many different types and forms, it is the metrics that relate to overall performance that should be of interest to you and your cloud solution. Trusting that everything is working to optimum performance is a sure-fire recipe for trouble down the road. You need to know it is. You need to monitor and watch it. And that’s where a great metrics suite becomes invaluable to you and your customers.
What to Monitor
Noam Alloush, the founder of SITE123, utilizes a simple acronym when it comes to valuable cloud metrics. PACE: performance, availability, capacity, and economics. Organizing and structuring your metrics around these four categories will ensure you’re keeping an eye on what is important to your users.
It’s hard to put a definitive definition on performance. What constitutes superior performance for some will fall well short for others. Everyone is different. But when it comes to the cloud, there are a few things that you’ll want to monitor, and none more important to most users than speed. Users want it fast. They want it virtually instantaneous, so your cloud product better be up to the task. Again, you could assume it is fast. You could take it for granted, but the metrics will provide concrete and trackable evidence. Is it as fast as other products? Is it getting faster over time (demonstrating quality control and improvement from your cloud partner), or is it slowing down?
Performance metrics might include input/output operations per second (IOPS), bandwidth (the speed of data transfer), response time, and latency (time for a particular operation to be completed). All of these give a “big picture” snapshot of overall speed, and that equates performance for most cloud users. In cars and cloud, fast = high performance.
And while speed is usually at the forefront from the end-user perspective, you might also want to consider reliability metrics as well. MTBF (mean time between failures) and MTTR (mean time to repair) metrics give evidence as to overall reliability. Are system failures few and far between, or are they increasing in frequency and proximity? Do failures and repairs take a lot of time to be addressed by your cloud partner? Reliability issues will drag your performance rating down, and that reflects poorly on you and your business (after all, in a white-label cloud product, it’s your logo on the platform, not your cloud partner’s).
Having files and documents in the cloud is both convenient and useful. Anywhere accessibility has increased productivity and efficiency across a multitude of industries, for both personal and professional tasks. But what about those (hopefully) rare instances when the system does not work as it is supposed to? Metrics that measure availability consider how accessible user files are in the event of a system failure or outage, and how often those problems occur.
The ideal in the tech world is what’s called “five 9s”. A 99.999% (five nines) availability rating means that there is five minutes and 15 seconds or LESS of total downtime during a year. Does your cloud partner deliver the five 9s? A rating of 99.99% looks equally impressive, but dropping those nine one-thousandths of a percent means nearly 53 minutes of downtime over the year. 99.9% jumps to eight hours and 46 minutes, while 99% is a staggering three days, 15 hours and 40 minutes of downtime per calendar year.
Monitor and demand consistency – if not improvement – from your partner. Availability metrics should also explore how many copies of data are stored, and how that affects accessibility in times of outage or failure.
Capacity metrics consider the cloud storage amount and will monitor how much storage customers are actually using compared to how much they have each been allocated.
Is the storage amount thin or thick/fat provisioned? Thick or fat provisioning is the physical allocation of storage space to a user. It is assigned and paid for, despite possibly never actually being used by the customer (someone using only 100MB of a 10GB account, for example). For thin provisioning, on the other hand, the space is utilized only as needed, and can be applied elsewhere when it’s not. The total capacity is shared among multiple users and allocated as required, allowing for maximum efficiency. With thin provisioning, that same individual using 100MB of a 10GB account is only using those 100MB. The rest of the storage is spread amongst other users requiring more space.
Simple dollars and cents. In any business, you need to evaluate cost at all times. What does something cost? What kind of return does that cost yield? Are there less expensive options out there? Your cloud solution shouldn’t be any different.
Metrics that collect economic data will likely examine energy and cost. They typically compare one set of data against another, and for that reason are often labeled compound metrics. The cost per gigabyte would be a good example, and well worth considering as part of your cloud solution metric suite. Keep costs down, or utilize the data to find ways to lower it.
How much storage is each user being provided? At what cost? How much of that storage is actually being used (these metrics come from the capacity category).
PACE. Performance, Availability, Capacity, and Economics. These are the four key areas that you should be watching. Your cloud metrics provide concrete data that can be tracked over time and used to improve each area in isolation, or any number of combinations.
A consumer cloud product is all about going above and beyond. Fall short in any section of PACE, and you either look bad to your users, or spend unnecessary money. Monitor, analyze, and use the data collected from your metrics to maximize your cloud offering. Ensure top-notch performance (speed and reliability) and your customers will spread the word (and your brand). Keep an eye on availability metrics and your users will believe their files are safe, secure, and accessible at virtually all times. If your cloud partner can’t promise that (and you’ll know if you’re watching the metric), then it’s time to find one who can.
Capacity and economic metrics go hand-in-hand for many things. You want a good bang for your buck, what you’re getting and for how much. It’s your name and logo on the cloud platform, so you’d better be sure of what is going on under the hood.