The cloud model of service-based IT delivery is a fabulous computing cure-all that provides unlimited no-fuss technology provision with no-hassle ongoing maintenance for an infinite period.
OK,[link widoczny dla zalogowanych], this obviously isn’t true, but what level of real application capacity control and resource allocation planning should firms actually use to corral and control their cloud deployments?
As pristine as any cloud migration (or new deployment) looks on paper, the ragged edges of real world data will leave any system unbalanced after a certain amount of time. A private cloud instance supporting 25 virtual machines (VMs) per server could, for example, be sitting on untapped hardware inefficiency where upping to 40 VMs per server is possible. But how do we know when to do this?
The Tetris technology template
CiRBA’s co-founder and SVP of product strategy Riyaz Somani describes the discovery process here as managing the economics of the operational envelope,[link widoczny dla zalogowanych]. Once a cloud deployment is in motion, Somani says that it’s a Tetris-like game of what fits best where depending on the shape and size of instances as they play out on a day to week to month basis.
You could almost call this process Windows disk defragmentation for clouds if you don’t like Tetris.
It’s a Tetris-like game of what fits best where, depending on the shape and size of instances,[link widoczny dla zalogowanych], as they play out on a day to week to month basis
The cloud application analytics software sold by CiRBA applies to the private cloud in the first instance, but at a higher level can also be used for public cloud instances where a customer uses a significant chunk of a datacentre.
“A workload could be located in separate datacentres A, B, C, or D and all would have different variables to consider. So cloud management advice needs to be given based upon individual SLAs,[link widoczny dla zalogowanych], latency, cost and overall performance,[link widoczny dla zalogowanych],” said Somani.
The theory extends thus. As software defined datacentres enjoy increasing prevalence and compose more of our overall network fabric, the physical geographic walls that make our data banks are broken down - and so more operational envelope economics can be potentially brought to bear.
Somani asserts that without his firm’s technology layers the cloud becomes comparatively inefficient - well, he would say that wouldn’t he?
However, the suggestion that cloud-based Virtual Machines are very often sold at the wrong size (or potentially at over capacity) does seem feasible. The rationale being that, until we know the history of managing a cloud workload, it is impossible to predict the provisioning requirement as accurately as we do post-deployment.
If we buy the CiRBA story,[link widoczny dla zalogowanych], then firms thinking about using the cloud for data analytics might need to do cloud analytics before data analytics. The firm’s technology works by defining policies to control what data tasks are carried out on what processing stacks in which cloud. A drag and drop functionality is offered in the newly released CiRBA Version 7.2 for virtual and cloud infrastructures.
This software is capable of analysing infrastructures ahead of cloud app migrations to visualise the impact of both hardware and software related changes on running environments, both now and in the future. Its cloud developer self-service application management if you like. CiRBA calls it “application transformation”, but lets try and keep it real.
"Self service is important, but perhaps even more important is the standardisation that cloud models bring, both from a capacity perspective and a configuration perspective," said Somani.
This technology also works pre-deployment says the firm. It can be used for pre-outsourcing (to cloud) optimisation workload planning. Workloads can be viewed based upon planned data flows prior to execution; they can then be reassessed on a daily (or longer timeframe) basis to see where applications should be potentially moved.
There’s no sizzling sausage to sell here, this is backroom cloud management for big customers like financial institutions. CiRBA’s Somani smiles and says hey, “It’s better than churning away at some ‘permutational engine’ for hours.”
Now there’s a snappy marketing catch line made in heaven,[link widoczny dla zalogowanych], right?

The post has been approved 0 times