The Dollars And Sense Of Business Continuity
September 21, 2020 Timothy Prickett Morgan
Those companies that have been in the IBM midrange market for a long time and are still alive and kicking have undergone a lot of change over the decades. But as is also the case, the core people at the company have been there for a long, long time and they understand how to leverage change to drive business and absorb change to help customers cope.
That is one of the secrets of longevity for Datanational, founded in September 1979, which is located in the Farmington Hills suburbs of Detroit and which has always had a strong presence in what was once called the Northwest Territories, what those of us who watch college football might call the Big Ten, and what some people referred to in the past as the Rust Belt. But, as the company’s name suggests, many of the clients that are located in this region have operations around the country, and so its business not only spans across decades, but spans the United States and extends across the globe.
We wanted to know a bit more about Datanational and the various services it offers, and we sat down with Derrick Smith, a business development manager at the company who specializes in its Business Continuity, System Protection and Recovery, and Data Protection offerings.
Timothy Prickett Morgan: Welcome to The Four Hundred. Tell us a little about yourself and then we will dig into what Datanational does here early in its fifth decade of existence.
Derrick Smith: I have been at the company since 1998, and I have a couple of different roles here and I have done a bunch of things during my career. I have worked with organizations to develop resiliency strategies that seamlessly responds to IT complexities and interdependencies across all cloud and on-premises resources. I enjoy helping customers to overcome challenges related to implementing different levels of business continuity, disaster recovery, and preventing disruptions from cyber attacks. We have a strong managed services offering here at Datanational and our Business Continuity Services is a pillar under that services offering. This has been a huge growth area for us – as you can imagine.
TPM: High availability has always been a key driver for the AS/400, iSeries, System i, and IBM i businesses, that is for sure, and I think we have learned there is elasticity of demand here in that as prices for systems and HA/DR software has come down, demand has gone up faster and the net-net is good growth.
Derrick Smith: My background since I was brought on board is much broader than this, of course. I learned everything you need to know about the AS/400, iSeries, System i, and IBM i infrastructure, down to the operating system and through down to the hardware platform, and then up to middleware and beyond. As I grew in my career, I was providing consulting services on how to build out redundancy within a customer’s environment. Back in the early late 1990s and early 2000s, providing an HA solution for a client cost a ton of money, and only the bigger IBM midrange shops could afford something like that. You needed a second system, a second location, and connectivity between those systems, and you needed expensive software to manage the replication. Now, obviously, all that has changed and we can give customers the closest thing to an “Easy” button as they can get. There is, of course, no real “Easy” button.
I have to listen very carefully to clients and discover what they really need from a recovery time objective and from a recovery point objective. Trying to determine what they really need is sometimes a real challenge. Everybody wants the Cadillac, but when it comes down to it, you know, sometimes they have a budget for a Chevy
TPM: I was talking to some CIOs and CTOs this week, and one of them said that this was a real issue within his company. And to make things more approachable, they created tiers like in a SkyMiles program, what I would call from Wood through Copper to Silver to Gold to Platinum to Uranium. No one wants wood – that’s tape backup – and no one can afford Uranium – that’s triple redundancy locally plus geographically distributed applications and databases across multiple, independent network backbones or whatever. He put a price on each HA level and then had them pick based on the interplay between the real needs of the application and their real budget constraints.
Derrick Smith: The challenge is that there are a mix of needs, and the mix is a higher number than the number of production systems and the number of logical partitions on the production system. Some parts of the business will need Platinum, using the scheme you laid out here, and some areas might be wood; some will be in-between. So, for instance, in a lot of cases, customers won’t want to spend a ton of money to support their human resources applications, which don’t provide a lot of differentiated value to the business. So they’re not going to want to invest in an HA solution or some type of time remote journaling solution for these parts.
We work with technology partners like Precisely, formerly Syncsort formerly Vision Solutions, and Carbonite, which is an OpenText company, to provide for a various forms of replication and vaulting to deliver to specific RTOs and RPOs. Through our partnerships, we deliver these solutions through a metered usage model where we have seen a big change in the last two years with logical replication of logical partitions based on the number of cores, where it is all metered. We can do that with Precisely’s MIMIX HA software, where we have worked out metered pricing for the software that can match the metered pricing we have on processing capacity in our business continuity center here in Farmington Hills. This allows us to be very competitive on pricing and implement a very affordable solution for clients.
Being located where we are, we are heavily involved in manufacturing and distribution, and back in the early days of the System/3X and AS/400 in the 1980s and 1990s, you had to have a value add to be a reseller, and ours was an ERP package called ACCESS/400. With our long standing heritage in the IBM i platform, we offer many services around ERP systems like he Epicor CMS, JD Edwards, Infor XA and Infor LX which run on IBM i and is serves the manufacturing and distribution industry very well. We also publish a solution called PlantTalk MES, which manages shop floors for manufacturers as the name suggests and which runs on Windows Server and SQL Server from Microsoft. We offer application support for many of these ERP applications and have a large number of developers and software engineers on staff. We also have customers in other parts of the economy, including retail and financial services. We do systems integration work in addition to offering managed services and HA/DR.
About 70 percent of our business overall at Datanational comes from the IBM i platform, with the other 30 percent coming from X86 server platforms – mostly Dell and Hewlett Packard Enterprise – running Windows Server and Linux. We can also do system integration for customers who have a mix of server types and who want to do high availability on Power Systems as well as X86 iron.
TPM: Let’s drill down on the high availability business a bit. Do you offer services based on MIMIX only, or can customers choose iTera HA as well?
Derrick Smith: We are certified to offer both, although the MIMIX solution allows for us to leverage the metered usage model making pricing pretty attractive and I can do it for one monthly cost for the HA solution, including the backup machine, the software, and the managed services all in one price. We have similar abilities for clients that are looking for Cloud Backup and Disaster Recovery where we would plug in the Carbonite technology.
For HA, A lot of IBM i customers do not want to provide their own hardware and simply want HA as a service – in fact, these days it is a large majority of clients. They’re sick of having hardware, and trying to keep up with the IBM i operating system releases and PTFs and hardware generations expiring. We don’t just provide the system but we have deep expertise in the replication software and we add on systems monitoring and management services on top of that and a staff that supports these machines 24 hours by 7 days by 365 days a year. Some IBM i customers are losing their experts to retirement, and they need help, and those that do not have IBM i expertise surely will not have MIMIX expertise.
TPM: If I came to Datanational and said I want to use Maxava HA or Robot HA or other disk-based replication tools, are customers prevented from doing that because of your strong partnership with Precisely?
Derrick Smith: No, absolutely not. And those are those are things that we have done in the past. So if someone has an existing implementation and they want to move it to a service, we can certainly do that. Or we can do it from scratch and there is nothing stopping us from doing that.
TPM: Let’s talk about pricing for a minute. Give me a rule of thumb here, just so I understand. Say I have 10,000 CPWs of oomph – that’s about one Power8 core running at about 3.5 GHz off the top of my head – in a logical partition that I want to replicate to your managed HA service. What is the range in price for that, all in, for the managed service and the capacity?
Derrick Smith: A ballpark price would be somewhere between $2,500 to $3,500 per month, depending on the level of services on it, such as active or passive monitoring. It would include the hardware, the systems software, the HA software, the management and monitoring services, and a 10 Mb/sec network link into the partition.
There is other stuff that we do in here, too, and it’s important. We get into these systems and look for things we need to be aware of in the application code, and there is some quirky stuff in there. For instance, sometimes a system throws out a message after the nightly processing is done, and if someone doesn’t reply to that, then the MRP system hangs. Weird stuff like that, it happens all the time. So we build a runbook for the HA system so we know all about the weird stuff and can deal with it for customers.
TPM: This seems like a pretty reasonable price for a replicated system. I like to think of this in terms of truck payments, so this is five to seven truck payments a month for a nice Dodge, Chevy, or Ford 4×4. I find it perplexing sometimes that people who work at companies with $10 million to $50 million in sales complain that a new IBM i machine costs $50,000 or $100,000 and yet it runs all of their mission critical applications for thousands or hundreds of thousands or millions of their own customers. And if they leased the IBM i machine, it would be one or two truck payments a month. Maybe three. We wouldn’t even think about having two cars as a person, and that is with several orders of magnitude less income than a company has.
Derrick Smith: A lot of our clients are not doing HA, but rather backup and recovery services, and for that, we partner with OpenText for its Carbonite EVault cloud backup service, and we restore to partitions we have running in our business continuity center. A lot of times we are working with customers who were just doing ridiculous things before moving to proper backups, like storing tape backups in the trunks of their cars or keeping them in their basement refrigerators.
TPM: The IBM i datacenter is not just about processing capacity embodied in Power Systems. Most companies also have X86 servers running Windows Server or Linux, and while the ratio between these has changed over the years, the tendency to offload to X86 iron or to augment the IBM i iron with X86 servers is not new. For many years, application servers for ERP suites that were on the OS/400 and then IBM i platform could run on X86 servers, and many customers did just that. In other cases, OLAP and other kinds of analytical processing was offloaded to Windows Server running SQL Server. Companies have to protect this iron and those applications, too.
Derrick Smith: With a lot of IBM i clients, we start with Carbonite EVault on the IBM i platform for backup and recovery, and they see this is working well and then they do it on their X86 servers. Then they move to HA/DR on IBM i in some cases where their uptime requirements are higher, and then some of them then want the same thing on their X86 servers. For that, we use Carbonite Availability, and like MIMIX, we can get metered pricing from OpenText and map the costs to the usage level so that works out.
TPM: What is the behavior across these customer sets, Power Systems-IBM i and X86-Windows/Linux? Who has just tape, who is doing EVault, who is doing HA? What is your estimate for the shape of the slices on this pair of pie charts?
Derrick Smith: In the IBM i base across our customers here at Datanational, there are probably about 20 percent that are using tapes still, with another 30 percent using EVault, 20 percent using EVault in combination with some kind of managed services, and the remaining 30 percent using true HA software. Across the X86 base among Datanational customers, where there are just so many competitive offerings that have been less expensive for longer, only about 10 percent of customers rely just in tape, with 40 percent using EVault in some fashion and the remaining 50 percent having some HA. This is often SAN replication, not just HA software like Carbonite Availability.
This content is sponsored by Datanational.