TECHNOSAVVIE: “Cloud And Hybrid Hosting Are The Way To Go!”

ezmob_search

adport_banner728x90

coinads

7search

zerads_728x90

multiwall_728x90

Search This Blog

Sunday, 23 February 2014

“Cloud And Hybrid Hosting Are The Way To Go!”

Struggling with the complexities in the SLA (Service Level Agreement) of a data centre service provider? Well, here’s some help. Today, when the cloud is the biggest phenomenon, it is important for IT admins to know what they are entitled to before inking an SLA. Diksha P Gupta from Open Source For You spoke to Anil Chandaliya, chief network administrator, ESDS, about how customers can play safe while dealing with data centres. Excerpts:

Q: In the data centre business, what are the latest trends, particularly with respect to SMBs?

If you talk about SMBs, there are many companies trying to get on the cloud. They are curious about it. There are generally two kinds of customers: the first type are the SMBs that have their own IT teams and their decision maker is the IT manager, while in the second model, the SMBs have consultants. If they resort to consultants, the firms tend to follow their advice and consultants generally don’t prefer the cloud or keeping the data in data centres. They still prefer keeping data within their premises because of certain inhibitions about storing data out of the office premises. The customers who have in-house IT managers are ready to take on different challenges and risks. They are IT savvy guys.

In the days to come, the adoption of the Cloud will surely increase. Also, hybrid hosting is one concept that is going to be a lot more prevalent for SMBs in times to come. In the case of hybrid hosting, there are many government organisations and enterprises that have already invested in on-premise data centres. Such organisations are trying to get on to the cloud because now they have realised that they get many advantages with it. One of the main reasons is that they have a lot of server resources that are unutilised. In which case, the cloud becomes a great solution. A hybrid cloud can meet their dynamic requirements and it can save them money as well. It allows them to have in-house virtualisation and whenever required, they can connect to a public cloud and draw resources from there. This scaling can be done for the time being, where they have to pay only for the resources they use. They can scale down after their requirements are met.





Q: Please share some details about the hybrid cloud and how secure it is.

IT admins are still not entirely convinced about the public cloud scenario, which is why hybrid clouds are coming into the picture. For example, in a bank set-up that has its database in its own premises, the bank may be using 10 different applications and putting load on application servers. What they can do, while keeping the data inside their premises, is to leverage the infrastructure of a third party data centre. So, on runtime, the data centre can offer the bank 10 virtual machines, which can be leveraged upon for a particular period, while keeping their data with themselves. In this way, the issues related to security are resolved easily, while they are leveraging the advantages of the public cloud. They are saved from increasing their infrastructure capacity, which would remain unutilised after some time. I think the hybrid cloud is here to stay.

Regarding the security aspects of a hybrid cloud, in the case of enterprise customers, we always have secure connectivity like Multiprotocol Label Switching (MPLS) or point-to-point connectivity. MPLS is basically a VPN between the network on the customers’ premises and commercial data centre. So all the communication that will happen between the customers’existing system, their database and the system they are getting from commercial data centre will be secure.

Q: What should an SMB look at before getting on a hybrid cloud?

Nowadays, it is not just about the service that the data centres provide; it is about the complete solution that they offer. So, before one chooses a cloud service provider, do ensure that what you’re signing on for is beyond the normal service offering and is a solution-based model.

The first thing that anyone should look at is a true pay-per-consume model. Users should have the flexibility to create a virtual machine for required time period and within that too have the flexibility to pay only for the actual resources being used out of the allocated resources. If you don’t use your virtual machine at night, you should not be charged for that period.
Second, along with the virtual machine, customers get a panel on which they can view how much RAM they use, as well as their utilisation of bandwidth and CPU. If a data centre’s system is dynamic, it will assess the periods of maximum load automatically and scale the resources, including RAM and CPU, on need basis without any human intervention.

Most importantly, customers should consider high availability. Ultimately, even if my virtual machine or server is running, but if my applications like the HTTP service, email service or Web service, are down, then the arrangement is of no use. So the uptime of applications is also very important.
The security of a cloud is a major concern for IT admins; so one should understand what security parameters have been met by the cloud service provider. Also, one should assess the quality of services being delivered. There should be a dedicated account manager for every enterprise customer. There should also be a proper SLA in place, because each of our customers will have their own customers and the account manager should help them immediately.

Q: What should one be looking for in the SLA?

The situation is such that whenever an SLA is put in place, everyone wants to be on the safe side. One major factor that should be kept in mind is the difference between the response time and the resolution time. Generally, in an SLA, the data centres talk about the response time and not the resolution time. Response time is when I call the data centre’s call centre and inform them about my issue, and they agree to get back to you in the stipulated time. Even if they get back within the said time, but they resolve the issue much later, it is a concern. They may keep you updated about the issue or keep responding in one way or the other, but will that resolve the issue? Most probably not. So customers should check what the resolution time mentioned in the SLA is and against that, what is the remedy agreed upon if they do not resolve the issue within the said time. The remedy is basically the penalty for not living up to the SLA and covering the losses that the customer may have faced due to a break in the services.

Yet another point that the customer should clarify is the definition of ‘downtime’ or ‘service outage’. If there is downtime, the customer may end up losing revenue but the data centre authorities may say that the service outage is not from their end. Both the parties need to agree that when the application is down, it should be considered as ‘service outage’. For the customer, the important part is the functioning of the application, while for the data centre, everything is restricted to the server and service. So, both customer and service provider should be on the same page.

Proactiveness from the customers is also required. They should look at all the points in the SLA and clearly understand them before getting into any agreement. 

No comments:

Post a Comment

ezmob_inpage

Featured post

What should students, parents, and teachers know about AI?

AI education will help people understand the risks, limitations, and opportunities Former judge Kay Firth-Butterfield began to think about h...

multiwall_300x250