Crawl. Walk. Run. That progression pretty well summarizes how most people look to take advantage of cloud service providers over time though, in cloud services terminology, the progression may be better summed up as: Archive, Replicate, Recover. Today I conclude my conversation with American Internet Service’s VP of Network Engineering, Steve Wallace, as we examine how many of AIS’ clients initially get their data into the AIS cloud and then expand their use of AIS cloud services over time.
Jerome: In terms of how you implement your offering, there are a number of ways to configure Nirvanix. Do you encourage people to deploy Nirvanix inside of their environment and virtualize their existing storage in pass thru mode so transactional data may be both kept local and moved offsite to either your data center or another one of their data centers? Alternatively, do you give them an appliance so they may do a mirrored write or make a copy of their existing data? Or do you give them multiple options?
Wallace: Nirvanix is an archival solution and is really not intended for live replication of transactional data. It is an excellent way to ensure that you have a recovery point that is independent of your local facilities. Their service is generally used with a gateway application or appliance that manages the ingestion of data into their storage nodes. For the folks who do not have the manpower or expertise to implement that type of solution, AIS provides a simple appliance.
We present them with an NFS or CIFS share that is on a VM, so they can essentially dump their data and have the appliance send it out into the cloud. That makes it a one step process since most folks understand the concept of a CIFS or NFS share and grasp that if they copy data or dump data on the share it will get to the archive.
Another AIS benefit is that we also operate a large wide-area network, so it costs us virtually nothing to send data to Nirvanix. We make sure the clients do not get charged for that data that gets pushed out.
Jerome: Generally the easiest way for people to move to the cloud is to provide them with a “bolt-on” type of approach. In that scenario, they leave their existing infrastructure in place but add a gateway, file share or portal to it that they can access. They then use their backup software, archival software, data management software – whatever -that copies their data to the portal that puts data into the cloud. Would that be a correct way to typify how you are encouraging people to move into the cloud?
Wallace: That is correct. When you say cloud in this sense, we are just talking about cloud archiving because we have cloud services, which in some ways, are actually easier to manage than that.
When looking at a DR scenario in a state like California, even if you push your data up to the Nirvanix cloud archive and California becomes a smoking crater – and you do not have infrastructure anywhere else you are in big trouble! It’s great, you have your data, but what are you going to do to get back online?
The DR solution is really incomplete without having some type of virtualized infrastructure and that is a service we provide. We can recreate your primary application infrastructure on a virtual infrastructure at our remote facility and then synchronize data between the two sites.
That pushes back the Nirvanix solution to a solution of a last resort or a protected archive service. If all else fails, we still have that. The data is fully protected and we know exactly where it is.
Jerome: So if I understand you correctly, you use Nirvanix as a cloud archive offering as a way to help organizations get started in the cloud. Once there, they are more comfortable starting this 360-degree discussion about how they can create a more highly available, recoverable environment for their applications within the state, out of the state or even internationally. Is that a correct conclusion to draw?
Wallace: That is correct. The first thing you need to do is get your data offsite. Get it somewhere else but somewhere other than on a USB drive that you put in the trunk of your car every week and swap out. Get that data somewhere else where it is accessible. That is a great play for the cloud archiving.
The next step is to get down to the practicalities of recovery. That is the next logical step, which is true even for smaller businesses. Then you are looking at recovering in hours as opposed to being down for a couple of weeks.
Jerome: So to summarize what AIS is doing: AIS gives companies an entry point to solve their immediate cloud archiving, cloud backup or compliance requirements. Then when a company is ready, it can take it up to this 360 degree approach.
Wallace: That is a good overview. 85% of our clients use some sort of cloud infrastructure and can extend their IT infrastructure into the AIS cloud. We have built our cloud infrastructure deliberately so it can be used as an extension of their existing cloud.
Our infrastructure is compatible with VMware, so they can manage it from their own space with their own management tools. It is totally private and secure. So rather than having to buy more equipment for their corporate datacenter, they can put a network cable in place and their cloud servers and our cloud servers will effectively sit right next to one another in the same rack.
That concept may be extended even to our Phoenix site. If they want to have a backup DR with their backup infrastructure ready to go, we can transmit their data across our 10 Gb links to Phoenix. We can achieve almost synchronous data replication across that link.
What we are hearing from our customers these days is that none of them want to buy any more equipment. We are their infrastructure service provider. We provide space, power and cooling so they don’t have to manage data centers. They recognize that is not their core business, so they expect us to do that for them. Now we talk about utility compute and utility storage – so they don’t have the expense and distraction of managing hardware.
That is how many cloud service providers in general and AIS specifically have evolved. We started by getting their data offsite and into the data center. Now that we have our client’s data offsite, we want to replicate it into another data center or another safe place. With the cloud, we can mirror their physical environment locally and in a remote data center with a much lower cost. We still need to keep that safety net of cloud archiving.
The economic drivers push businesses to reduce the total cost of ownership. No one wants to own and maintain cloud archiving, storage or any other type of storage. It is not as simple as you would think and the data requirements are tremendous.
The genomics people sometimes require 100 TBs or more for raw data. After they are done processing it, it might drop down to a TB or two but considering the amount of jobs they are processing, we are still talking huge amounts of data. That’s a lot of infrastructure built around moving and storing data. These clients are not interested in owning that infrastructure – they’d rather be sequencing genomes.
In Part I of this interview series with Steve Wallace, we talk about how a convergence in service offerings is occurring among cloud service providers driven by their client needs for them to offer business continuity, disaster recovery and compliance services.